kernel/linux
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ synced 2026-11-15 19:56:03 +00:00
Code Issues Releases Packages Activity
linux/block/blk-iolatency.c

1068 lines
29 KiB
Raw Normal View History

block: add SPDX tags to block layer files missing licensing information Various block layer files do not have any licensing information at all. Add SPDX tags for the default kernel GPLv2 license to those. Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-04-30 14:42:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: fix max_depth comparisons max_depth used to be a u64, but I changed it to a unsigned int but didn't convert my comparisons over everywhere. Fix by using UINT_MAX everywhere instead of (u64)-1. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-11 10:34:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: make iolatency avg_lat exponentially decay Currently, avg_lat is calculated by accumulating the mean of every window in a long running cumulative average. As time goes on, the metric becomes less and less useful due to the accumulated history. This patch reuses the same calculation done in load averages to make the avg_lat metric more lively. Unlike load averages, the avg only advances when a window elapses (due to an io). Idle periods extend the most recent window. Bucketing is used to limit the history of avg_lat by binding it to the window size. So, the window range for 1/exp (decay rate) is [1 min, 2.5 min) when windows elapse immediately. The current sample window size is exposed in the debug info to enable calculation of the window range. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-08-01 23:15:41 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: fix IO hang due to negative inflight counter Our test reported the following stack, and vmcore showed that ->inflight counter is -1. [ffffc9003fcc38d0] __schedule at ffffffff8173d95d [ffffc9003fcc3958] schedule at ffffffff8173de26 [ffffc9003fcc3970] io_schedule at ffffffff810bb6b6 [ffffc9003fcc3988] blkcg_iolatency_throttle at ffffffff813911cb [ffffc9003fcc3a20] rq_qos_throttle at ffffffff813847f3 [ffffc9003fcc3a48] blk_mq_make_request at ffffffff8137468a [ffffc9003fcc3b08] generic_make_request at ffffffff81368b49 [ffffc9003fcc3b68] submit_bio at ffffffff81368d7d [ffffc9003fcc3bb8] ext4_io_submit at ffffffffa031be00 [ext4] [ffffc9003fcc3c00] ext4_writepages at ffffffffa03163de [ext4] [ffffc9003fcc3d68] do_writepages at ffffffff811c49ae [ffffc9003fcc3d78] __filemap_fdatawrite_range at ffffffff811b6188 [ffffc9003fcc3e30] filemap_write_and_wait_range at ffffffff811b6301 [ffffc9003fcc3e60] ext4_sync_file at ffffffffa030cee8 [ext4] [ffffc9003fcc3ea8] vfs_fsync_range at ffffffff8128594b [ffffc9003fcc3ee8] do_fsync at ffffffff81285abd [ffffc9003fcc3f18] sys_fsync at ffffffff81285d50 [ffffc9003fcc3f28] do_syscall_64 at ffffffff81003c04 [ffffc9003fcc3f50] entry_SYSCALL_64_after_swapgs at ffffffff81742b8e The ->inflight counter may be negative (-1) if 1) blk-iolatency was disabled when the IO was issued, 2) blk-iolatency was enabled before this IO reached its endio, 3) the ->inflight counter is decreased from 0 to -1 in endio() In fact the hang can be easily reproduced by the below script, H=/sys/fs/cgroup/unified/ P=/sys/fs/cgroup/unified/test echo "+io" > $H/cgroup.subtree_control mkdir -p $P echo $$ > $P/cgroup.procs xfs_io -f -d -c "pwrite 0 4k" /dev/sdg echo "`cat /sys/block/sdg/dev` target=1000000" > $P/io.latency xfs_io -f -d -c "pwrite 0 4k" /dev/sdg This fixes the problem by freezing the queue so that while enabling/disabling iolatency, there is no inflight rq running. Note that quiesce_queue is not needed as this only updating iolatency configuration about which dispatching request_queue doesn't care. Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-01-25 08:12:47 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: partition include/linux/blk-cgroup.h Partition include/linux/blk-cgroup.h into two parts: one is public part, the other is block layer private part. Suggested by Christoph Hellwig. Signed-off-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20220211101149.2368042-4-ming.lei@redhat.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-02-11 18:11:49 +08:00
blk-iolatency: #include "blk.h" This patch avoids that the following warning is reported when building with W=1: block/blk-iolatency.c:734:5: warning: no previous prototype for 'blk_iolatency_init' [-Wmissing-prototypes] Cc: Josef Bacik <jbacik@fb.com> Fixes: d70675121546 ("block: introduce blk-iolatency io controller") # v4.19 Signed-off-by: Bart Van Assche <bvanassche@acm.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-03-20 13:15:01 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: make iolatency avg_lat exponentially decay Currently, avg_lat is calculated by accumulating the mean of every window in a long running cumulative average. As time goes on, the metric becomes less and less useful due to the accumulated history. This patch reuses the same calculation done in load averages to make the avg_lat metric more lively. Unlike load averages, the avg only advances when a window elapses (due to an io). Idle periods extend the most recent window. Bucketing is used to limit the history of avg_lat by binding it to the window size. So, the window range for 1/exp (decay rate) is [1 min, 2.5 min) when windows elapse immediately. The current sample window size is exposed in the debug info to enable calculation of the window range. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-08-01 23:15:41 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: make iolatency avg_lat exponentially decay Currently, avg_lat is calculated by accumulating the mean of every window in a long running cumulative average. As time goes on, the metric becomes less and less useful due to the accumulated history. This patch reuses the same calculation done in load averages to make the avg_lat metric more lively. Unlike load averages, the avg only advances when a window elapses (due to an io). Idle periods extend the most recent window. Bucketing is used to limit the history of avg_lat by binding it to the window size. So, the window range for 1/exp (decay rate) is [1 min, 2.5 min) when windows elapse immediately. The current sample window size is exposed in the debug info to enable calculation of the window range. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-08-01 23:15:41 -07:00
sched: loadavg: consolidate LOAD_INT, LOAD_FRAC, CALC_LOAD There are several definitions of those functions/macros in places that mess with fixed-point load averages. Provide an official version. [akpm@linux-foundation.org: fix missed conversion in block/blk-iolatency.c] Link: http://lkml.kernel.org/r/20180828172258.3185-5-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Tested-by: Suren Baghdasaryan <surenb@google.com> Tested-by: Daniel Drake <drake@endlessm.com> Cc: Christopher Lameter <cl@linux.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <jweiner@fb.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Enderborg <peter.enderborg@sony.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vinayak Menon <vinmenon@codeaurora.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-26 15:06:11 -07:00
block: make iolatency avg_lat exponentially decay Currently, avg_lat is calculated by accumulating the mean of every window in a long running cumulative average. As time goes on, the metric becomes less and less useful due to the accumulated history. This patch reuses the same calculation done in load averages to make the avg_lat metric more lively. Unlike load averages, the avg only advances when a window elapses (due to an io). Idle periods extend the most recent window. Bucketing is used to limit the history of avg_lat by binding it to the window size. So, the window range for 1/exp (decay rate) is [1 min, 2.5 min) when windows elapse immediately. The current sample window size is exposed in the debug info to enable calculation of the window range. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-08-01 23:15:41 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
sched: loadavg: consolidate LOAD_INT, LOAD_FRAC, CALC_LOAD There are several definitions of those functions/macros in places that mess with fixed-point load averages. Provide an official version. [akpm@linux-foundation.org: fix missed conversion in block/blk-iolatency.c] Link: http://lkml.kernel.org/r/20180828172258.3185-5-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Tested-by: Suren Baghdasaryan <surenb@google.com> Tested-by: Daniel Drake <drake@endlessm.com> Cc: Christopher Lameter <cl@linux.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <jweiner@fb.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Enderborg <peter.enderborg@sony.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vinayak Menon <vinmenon@codeaurora.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-26 15:06:11 -07:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
sched: loadavg: consolidate LOAD_INT, LOAD_FRAC, CALC_LOAD There are several definitions of those functions/macros in places that mess with fixed-point load averages. Provide an official version. [akpm@linux-foundation.org: fix missed conversion in block/blk-iolatency.c] Link: http://lkml.kernel.org/r/20180828172258.3185-5-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Tested-by: Suren Baghdasaryan <surenb@google.com> Tested-by: Daniel Drake <drake@endlessm.com> Cc: Christopher Lameter <cl@linux.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <jweiner@fb.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Peter Enderborg <peter.enderborg@sony.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vinayak Menon <vinmenon@codeaurora.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-26 15:06:11 -07:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: convert io-latency to use rq_qos_wait Now that we have this common helper, convert io-latency over to use it as well. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-04 12:59:04 -05:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: convert io-latency to use rq_qos_wait Now that we have this common helper, convert io-latency over to use it as well. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-04 12:59:04 -05:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: convert io-latency to use rq_qos_wait Now that we have this common helper, convert io-latency over to use it as well. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-04 12:59:04 -05:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: remove the unused lock argument to rq_qos_throttle Unused now that the legacy request path is gone. Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-11-14 17:02:09 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-rq-qos: store a gendisk instead of request_queue in struct rq_qos This is what about half of the users already want, and it's only going to grow more. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-16-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:56 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: convert io-latency to use rq_qos_wait Now that we have this common helper, convert io-latency over to use it as well. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-04 12:59:04 -05:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-rq-qos: store a gendisk instead of request_queue in struct rq_qos This is what about half of the users already want, and it's only going to grow more. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-16-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:56 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Correct comment for scale_cookie_change Default queue depth of iolatency_grp is unlimited, so we scale down quickly(once by half) in scale_cookie_change. Remove the "subtract 1/16th" part which is not the truth and add the actual way we scale down. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Link: https://lore.kernel.org/r/20221018111240.22612-3-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:39 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Correct comment for scale_cookie_change Default queue depth of iolatency_grp is unlimited, so we scale down quickly(once by half) in scale_cookie_change. Remove the "subtract 1/16th" part which is not the truth and add the actual way we scale down. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Link: https://lore.kernel.org/r/20221018111240.22612-3-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:39 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-rq-qos: store a gendisk instead of request_queue in struct rq_qos This is what about half of the users already want, and it's only going to grow more. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-16-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:56 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: deal with nr_requests == 1 Hitting the case where blk_queue_depth() returned 1 uncovered the fact that iolatency doesn't actually handle this case properly, it simply doesn't scale down anybody. For this case we should go straight into applying the time delay, which we weren't doing. Since we already limit the floor at 1 request this if statement is not needed, and this allows us to set our depth to 1 which allows us to apply the delay if needed. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:40 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Use atomic{,64}_try_cmpxchg Use atomic_try_cmpxchg instead of atomic_cmpxchg (*ptr, old, new) == old in check_scale_change and atomic64_try_cmpxchg in blkcg_iolatency_done_bio. x86 CMPXCHG instruction returns success in ZF flag, so this change saves a compare after cmpxchg (and related move instruction in front of cmpxchg). No functional change intended. Signed-off-by: Uros Bizjak <ubizjak@gmail.com> Cc: Jens Axboe <axboe@kernel.dk> Link: https://lore.kernel.org/r/20220712151947.6783-1-ubizjak@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-12 17:19:47 +02:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Use atomic{,64}_try_cmpxchg Use atomic_try_cmpxchg instead of atomic_cmpxchg (*ptr, old, new) == old in check_scale_change and atomic64_try_cmpxchg in blkcg_iolatency_done_bio. x86 CMPXCHG instruction returns success in ZF flag, so this change saves a compare after cmpxchg (and related move instruction in front of cmpxchg). No functional change intended. Signed-off-by: Uros Bizjak <ubizjak@gmail.com> Cc: Jens Axboe <axboe@kernel.dk> Link: https://lore.kernel.org/r/20220712151947.6783-1-ubizjak@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-12 17:19:47 +02:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: deal with small samples There is logic to keep cgroups that haven't done a lot of IO in the most recent scale window from being punished for over-active higher priority groups. However for things like ssd's where the windows are pretty short we'll end up with small numbers of samples, so 5% of samples will come out to 0 if there aren't enough. Make the floor 1 sample to keep us from improperly bailing out of scaling down. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:41 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: remove the unused lock argument to rq_qos_throttle Unused now that the legacy request path is gone. Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-11-14 17:02:09 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: associate blkg when associating a device Previously, blkg association was handled by controller specific code in blk-throttle and blk-iolatency. However, because a blkg represents a relationship between a blkcg and a request_queue, it makes sense to keep the blkg->q and bio->bi_disk->queue consistent. This patch moves association into the bio_set_dev macro(). This should cover the majority of cases where the device is set/changed keeping the two pointers consistent. Fallback code is added to blkcg_bio_issue_check() to catch any missing paths. Signed-off-by: Dennis Zhou <dennis@kernel.org> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-05 12:10:31 -05:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: remove the unused lock argument to rq_qos_throttle Unused now that the legacy request path is gone. Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-11-14 17:02:09 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: cleanup bio_issue Now that bio->bi_issue is only used by blk-iolatency to get bio issue time, replace bio_issue with u64 time directly and remove bio_issue to make code cleaner. Signed-off-by: Yu Kuai <yukuai3@huawei.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-09-10 14:30:41 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: only call ktime_get() if needed ktime_to_ns(ktime_get()), which is expensive, does not need to be called if blk_iolatency_enabled() return false in blkcg_iolatency_done_bio(). Postponing ktime_to_ns(ktime_get()) execution reduces the CPU usage when blk_iolatency is disabled. Signed-off-by: Hongnan Li <hongnan.li@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-07-01 16:09:38 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
Blk-iolatency: warn on negative inflight IO counter This is to catch any unexpected negative value of inflight IO counter. Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-01-25 08:12:48 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: fix rq-qos breakage from skipping rq_qos_done_bio() a647a524a467 ("block: don't call rq_qos_ops->done_bio if the bio isn't tracked") made bio_endio() skip rq_qos_done_bio() if BIO_TRACKED is not set. While this fixed a potential oops, it also broke blk-iocost by skipping the done_bio callback for merged bios. Before, whether a bio goes through rq_qos_throttle() or rq_qos_merge(), rq_qos_done_bio() would be called on the bio on completion with BIO_TRACKED distinguishing the former from the latter. rq_qos_done_bio() is not called for bios which wenth through rq_qos_merge(). This royally confuses blk-iocost as the merged bios never finish and are considered perpetually in-flight. One reliably reproducible failure mode is an intermediate cgroup geting stuck active preventing its children from being activated due to the leaf-only rule, leading to loss of control. The following is from resctl-bench protection scenario which emulates isolating a web server like workload from a memory bomb run on an iocost configuration which should yield a reasonable level of protection. # cat /sys/block/nvme2n1/device/model Samsung SSD 970 PRO 512GB # cat /sys/fs/cgroup/io.cost.model 259:0 ctrl=user model=linear rbps=834913556 rseqiops=93622 rrandiops=102913 wbps=618985353 wseqiops=72325 wrandiops=71025 # cat /sys/fs/cgroup/io.cost.qos 259:0 enable=1 ctrl=user rpct=95.00 rlat=18776 wpct=95.00 wlat=8897 min=60.00 max=100.00 # resctl-bench -m 29.6G -r out.json run protection::scenario=mem-hog,loops=1 ... Memory Hog Summary ================== IO Latency: R p50=242u:336u/2.5m p90=794u:1.4m/7.5m p99=2.7m:8.0m/62.5m max=8.0m:36.4m/350m W p50=221u:323u/1.5m p90=709u:1.2m/5.5m p99=1.5m:2.5m/9.5m max=6.9m:35.9m/350m Isolation and Request Latency Impact Distributions: min p01 p05 p10 p25 p50 p75 p90 p95 p99 max mean stdev isol% 15.90 15.90 15.90 40.05 57.24 59.07 60.01 74.63 74.63 90.35 90.35 58.12 15.82 lat-imp% 0 0 0 0 0 4.55 14.68 15.54 233.5 548.1 548.1 53.88 143.6 Result: isol=58.12:15.82% lat_imp=53.88%:143.6 work_csv=100.0% missing=3.96% The isolation result of 58.12% is close to what this device would show without any IO control. Fix it by introducing a new flag BIO_QOS_MERGED to mark merged bios and calling rq_qos_done_bio() on them too. For consistency and clarity, rename BIO_TRACKED to BIO_QOS_THROTTLED. The flag checks are moved into rq_qos_done_bio() so that it's next to the code paths that set the flags. With the patch applied, the above same benchmark shows: # resctl-bench -m 29.6G -r out.json run protection::scenario=mem-hog,loops=1 ... Memory Hog Summary ================== IO Latency: R p50=123u:84.4u/985u p90=322u:256u/2.5m p99=1.6m:1.4m/9.5m max=11.1m:36.0m/350m W p50=429u:274u/995u p90=1.7m:1.3m/4.5m p99=3.4m:2.7m/11.5m max=7.9m:5.9m/26.5m Isolation and Request Latency Impact Distributions: min p01 p05 p10 p25 p50 p75 p90 p95 p99 max mean stdev isol% 84.91 84.91 89.51 90.73 92.31 94.49 96.36 98.04 98.71 100.0 100.0 94.42 2.81 lat-imp% 0 0 0 0 0 2.81 5.73 11.11 13.92 17.53 22.61 4.10 4.68 Result: isol=94.42:2.81% lat_imp=4.10%:4.68 work_csv=58.34% missing=0% Signed-off-by: Tejun Heo <tj@kernel.org> Fixes: a647a524a467 ("block: don't call rq_qos_ops->done_bio if the bio isn't tracked") Cc: stable@vger.kernel.org # v5.15+ Cc: Ming Lei <ming.lei@redhat.com> Cc: Yu Kuai <yukuai3@huawei.com> Reviewed-by: Ming Lei <ming.lei@redhat.com> Link: https://lore.kernel.org/r/Yi7rdrzQEHjJLGKB@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-03-13 21:15:02 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-iolatency: fix IO hang due to negative inflight counter Our test reported the following stack, and vmcore showed that ->inflight counter is -1. [ffffc9003fcc38d0] __schedule at ffffffff8173d95d [ffffc9003fcc3958] schedule at ffffffff8173de26 [ffffc9003fcc3970] io_schedule at ffffffff810bb6b6 [ffffc9003fcc3988] blkcg_iolatency_throttle at ffffffff813911cb [ffffc9003fcc3a20] rq_qos_throttle at ffffffff813847f3 [ffffc9003fcc3a48] blk_mq_make_request at ffffffff8137468a [ffffc9003fcc3b08] generic_make_request at ffffffff81368b49 [ffffc9003fcc3b68] submit_bio at ffffffff81368d7d [ffffc9003fcc3bb8] ext4_io_submit at ffffffffa031be00 [ext4] [ffffc9003fcc3c00] ext4_writepages at ffffffffa03163de [ext4] [ffffc9003fcc3d68] do_writepages at ffffffff811c49ae [ffffc9003fcc3d78] __filemap_fdatawrite_range at ffffffff811b6188 [ffffc9003fcc3e30] filemap_write_and_wait_range at ffffffff811b6301 [ffffc9003fcc3e60] ext4_sync_file at ffffffffa030cee8 [ext4] [ffffc9003fcc3ea8] vfs_fsync_range at ffffffff8128594b [ffffc9003fcc3ee8] do_fsync at ffffffff81285abd [ffffc9003fcc3f18] sys_fsync at ffffffff81285d50 [ffffc9003fcc3f28] do_syscall_64 at ffffffff81003c04 [ffffc9003fcc3f50] entry_SYSCALL_64_after_swapgs at ffffffff81742b8e The ->inflight counter may be negative (-1) if 1) blk-iolatency was disabled when the IO was issued, 2) blk-iolatency was enabled before this IO reached its endio, 3) the ->inflight counter is decreased from 0 to -1 in endio() In fact the hang can be easily reproduced by the below script, H=/sys/fs/cgroup/unified/ P=/sys/fs/cgroup/unified/test echo "+io" > $H/cgroup.subtree_control mkdir -p $P echo $$ > $P/cgroup.procs xfs_io -f -d -c "pwrite 0 4k" /dev/sdg echo "`cat /sys/block/sdg/dev` target=1000000" > $P/io.latency xfs_io -f -d -c "pwrite 0 4k" /dev/sdg This fixes the problem by freezing the queue so that while enabling/disabling iolatency, there is no inflight rq running. Note that quiesce_queue is not needed as this only updating iolatency configuration about which dispatching request_queue doesn't care. Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-01-25 08:12:47 +08:00
block: add blk_time_get_ns() and blk_time_get() helpers Convert any user of ktime_get_ns() to use blk_time_get_ns(), and ktime_get() to blk_time_get(), so we have a unified API for querying the current time in nanoseconds or as ktime. No functional changes intended, this patch just wraps ktime_get_ns() and ktime_get() with a block helper. Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-15 14:45:07 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
Blk-iolatency: warn on negative inflight IO counter This is to catch any unexpected negative value of inflight IO counter. Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-01-25 08:12:48 +08:00
blk-iolatency: fix STS_AGAIN handling The iolatency controller is based on rq_qos. It increments on rq_qos_throttle() and decrements on either rq_qos_cleanup() or rq_qos_done_bio(). a3fb01ba5af0 fixes the double accounting issue where blk_mq_make_request() may call both rq_qos_cleanup() and rq_qos_done_bio() on REQ_NO_WAIT. So checking STS_AGAIN prevents the double decrement. The above works upstream as the only way we can get STS_AGAIN is from blk_mq_get_request() failing. The STS_AGAIN handling isn't a real problem as bio_endio() skipping only happens on reserved tag allocation failures which can only be caused by driver bugs and already triggers WARN. However, the fix creates a not so great dependency on how STS_AGAIN can be propagated. Internally, we (Facebook) carry a patch that kills read ahead if a cgroup is io congested or a fatal signal is pending. This combined with chained bios progagate their bi_status to the parent is not already set can can cause the parent bio to not clean up properly even though it was successful. This consequently leaks the inflight counter and can hang all IOs under that blkg. To nip the adverse interaction early, this removes the rq_qos_cleanup() callback in iolatency in favor of cleaning up always on the rq_qos_done_bio() path. Fixes: a3fb01ba5af0 ("blk-iolatency: only account submitted bios") Debugged-by: Tejun Heo <tj@kernel.org> Debugged-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Dennis Zhou <dennis@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-07-05 17:09:09 -04:00
block: cleanup bio_issue Now that bio->bi_issue is only used by blk-iolatency to get bio issue time, replace bio_issue with u64 time directly and remove bio_issue to make code cleaner. Signed-off-by: Yu Kuai <yukuai3@huawei.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-09-10 14:30:41 +08:00
blk-iolatency: fix STS_AGAIN handling The iolatency controller is based on rq_qos. It increments on rq_qos_throttle() and decrements on either rq_qos_cleanup() or rq_qos_done_bio(). a3fb01ba5af0 fixes the double accounting issue where blk_mq_make_request() may call both rq_qos_cleanup() and rq_qos_done_bio() on REQ_NO_WAIT. So checking STS_AGAIN prevents the double decrement. The above works upstream as the only way we can get STS_AGAIN is from blk_mq_get_request() failing. The STS_AGAIN handling isn't a real problem as bio_endio() skipping only happens on reserved tag allocation failures which can only be caused by driver bugs and already triggers WARN. However, the fix creates a not so great dependency on how STS_AGAIN can be propagated. Internally, we (Facebook) carry a patch that kills read ahead if a cgroup is io congested or a fatal signal is pending. This combined with chained bios progagate their bi_status to the parent is not already set can can cause the parent bio to not clean up properly even though it was successful. This consequently leaks the inflight counter and can hang all IOs under that blkg. To nip the adverse interaction early, this removes the rq_qos_cleanup() callback in iolatency in favor of cleaning up always on the rq_qos_done_bio() path. Fixes: a3fb01ba5af0 ("blk-iolatency: only account submitted bios") Debugged-by: Tejun Heo <tj@kernel.org> Debugged-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Dennis Zhou <dennis@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-07-05 17:09:09 -04:00
blk-iolatency: Use atomic{,64}_try_cmpxchg Use atomic_try_cmpxchg instead of atomic_cmpxchg (*ptr, old, new) == old in check_scale_change and atomic64_try_cmpxchg in blkcg_iolatency_done_bio. x86 CMPXCHG instruction returns success in ZF flag, so this change saves a compare after cmpxchg (and related move instruction in front of cmpxchg). No functional change intended. Signed-off-by: Uros Bizjak <ubizjak@gmail.com> Cc: Jens Axboe <axboe@kernel.dk> Link: https://lore.kernel.org/r/20220712151947.6783-1-ubizjak@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-12 17:19:47 +02:00
blk-iolatency: fix STS_AGAIN handling The iolatency controller is based on rq_qos. It increments on rq_qos_throttle() and decrements on either rq_qos_cleanup() or rq_qos_done_bio(). a3fb01ba5af0 fixes the double accounting issue where blk_mq_make_request() may call both rq_qos_cleanup() and rq_qos_done_bio() on REQ_NO_WAIT. So checking STS_AGAIN prevents the double decrement. The above works upstream as the only way we can get STS_AGAIN is from blk_mq_get_request() failing. The STS_AGAIN handling isn't a real problem as bio_endio() skipping only happens on reserved tag allocation failures which can only be caused by driver bugs and already triggers WARN. However, the fix creates a not so great dependency on how STS_AGAIN can be propagated. Internally, we (Facebook) carry a patch that kills read ahead if a cgroup is io congested or a fatal signal is pending. This combined with chained bios progagate their bi_status to the parent is not already set can can cause the parent bio to not clean up properly even though it was successful. This consequently leaks the inflight counter and can hang all IOs under that blkg. To nip the adverse interaction early, this removes the rq_qos_cleanup() callback in iolatency in favor of cleaning up always on the rq_qos_done_bio() path. Fixes: a3fb01ba5af0 ("blk-iolatency: only account submitted bios") Debugged-by: Tejun Heo <tj@kernel.org> Debugged-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Dennis Zhou <dennis@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-07-05 17:09:09 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
treewide: Convert del_timer*() to timer_shutdown*() Due to several bugs caused by timers being re-armed after they are shutdown and just before they are freed, a new state of timers was added called "shutdown". After a timer is set to this state, then it can no longer be re-armed. The following script was run to find all the trivial locations where del_timer() or del_timer_sync() is called in the same function that the object holding the timer is freed. It also ignores any locations where the timer->function is modified between the del_timer*() and the free(), as that is not considered a "trivial" case. This was created by using a coccinelle script and the following commands: $ cat timer.cocci @@ expression ptr, slab; identifier timer, rfield; @@ ( - del_timer(&ptr->timer); + timer_shutdown(&ptr->timer); | - del_timer_sync(&ptr->timer); + timer_shutdown_sync(&ptr->timer); ) ... when strict when != ptr->timer ( kfree_rcu(ptr, rfield); | kmem_cache_free(slab, ptr); | kfree(ptr); ) $ spatch timer.cocci . > /tmp/t.patch $ patch -p1 < /tmp/t.patch Link: https://lore.kernel.org/lkml/20221123201306.823305113@linutronix.de/ Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Acked-by: Pavel Machek <pavel@ucw.cz> [ LED ] Acked-by: Kalle Valo <kvalo@kernel.org> [ wireless ] Acked-by: Paolo Abeni <pabeni@redhat.com> [ networking ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-12-20 13:45:19 -05:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-cgroup: pass a gendisk to blkcg_{de,}activate_policy Prepare for storing the blkcg information in the gendisk instead of the request_queue. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-17-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:57 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-rq-qos: constify rq_qos_ops These op vectors are constant, so mark them const. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-15-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:55 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
treewide, timers: Rename from_timer() to timer_container_of() Move this API to the canonical timer_*() namespace. [ tglx: Redone against pre rc1 ] Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/all/aB2X0jCKQO56WdMt@gmail.com
2025-05-09 07:51:14 +02:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: add blk_time_get_ns() and blk_time_get() helpers Convert any user of ktime_get_ns() to use blk_time_get_ns(), and ktime_get() to blk_time_get(), so we have a unified API for querying the current time in nanoseconds or as ktime. No functional changes intended, this patch just wraps ktime_get_ns() and ktime_get() with a block helper. Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-15 14:45:07 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
Revert "blk-cgroup: move the cgroup information to struct gendisk" This reverts commit 3f13ab7c80fdb0ada86a8e3e818960bc1ccbaa59 as a patch it depends on caused a few problems. Signed-off-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230214183308.1658775-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-14 19:33:04 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: rename blkg_try_get() to blkg_tryget() blkg reference counting now uses percpu_ref rather than atomic_t. Let's make this consistent with css_tryget. This renames blkg_try_get to blkg_tryget and now returns a bool rather than the blkg or %NULL. Signed-off-by: Dennis Zhou <dennis@kernel.org> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-12-05 12:10:39 -05:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: fix blkg leak in timer_fn At this point we have a ref on the blkg, we need to drop it if we don't have a iolat. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-31 12:39:02 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-mq: add QUEUE_FLAG_BIO_ISSUE_TIME bio->issue_time_ns is initialized for every bio, however, it's only used by blk-iolatency. Add a new queue_flag and only set this flag when blk-iolatency is enabled, so that extra blk_time_get_ns() can be saved for disks that blk-iolatency is not enabled. Signed-off-by: Yu Kuai <yukuai3@huawei.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-09-10 14:30:43 +08:00
block: force noio scope in blk_mq_freeze_queue When block drivers or the core block code perform allocations with a frozen queue, this could try to recurse into the block device to reclaim memory and deadlock. Thus all allocations done by a process that froze a queue need to be done without __GFP_IO and __GFP_FS. Instead of tying to track all of them down, force a noio scope as part of freezing the queue. Note that nvme is a bit of a mess here due to the non-owner freezes, and they will be addressed separately. Signed-off-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20250131120352.1315351-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-01-31 13:03:47 +01:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-mq: add QUEUE_FLAG_BIO_ISSUE_TIME bio->issue_time_ns is initialized for every bio, however, it's only used by blk-iolatency. Add a new queue_flag and only set this flag when blk-iolatency is enabled, so that extra blk_time_get_ns() can be saved for disks that blk-iolatency is not enabled. Signed-off-by: Yu Kuai <yukuai3@huawei.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-09-10 14:30:43 +08:00
block: force noio scope in blk_mq_freeze_queue When block drivers or the core block code perform allocations with a frozen queue, this could try to recurse into the block device to reclaim memory and deadlock. Thus all allocations done by a process that froze a queue need to be done without __GFP_IO and __GFP_FS. Instead of tying to track all of them down, force a noio scope as part of freezing the queue. Note that nvme is a bit of a mess here due to the non-owner freezes, and they will be addressed separately. Signed-off-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20250131120352.1315351-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2025-01-31 13:03:47 +01:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-iolatency: Make initialization lazy Other rq_qos policies such as wbt and iocost are lazy-initialized when they are configured for the first time for the device but iolatency is initialized unconditionally from blkcg_init_disk() during gendisk init. Lazy init is beneficial because rq_qos policies add runtime overhead when initialized as every IO has to walk all registered rq_qos callbacks. This patch switches iolatency to lazy initialization too so that it only registered its rq_qos policy when it is first configured. Note that there is a known race condition between blkcg config file writes and del_gendisk() and this patch makes iolatency susceptible to it by exposing the init path to race against the deletion path. However, that problem already exists in iocost and is being worked on. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20230413000649.115785-5-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:49 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
Convert 'alloc_obj' family to use the new default GFP_KERNEL argument This was done entirely with mindless brute force, using git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' | xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/' to convert the new alloc_obj() users that had a simple GFP_KERNEL argument to just drop that argument. Note that due to the extreme simplicity of the scripting, any slightly more complex cases spread over multiple lines would not be triggered: they definitely exist, but this covers the vast bulk of the cases, and the resulting diff is also then easier to check automatically. For the same reason the 'flex' versions will be done as a separate conversion. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2026-02-21 16:37:42 -08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-rq-qos: make rq_qos_add and rq_qos_del more useful Switch to passing a gendisk, and make rq_qos_add initialize all required fields and drop the not required q argument from rq_qos_del. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-14-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:54 +01:00
block: don't allow the same type rq_qos add more than once In our test of iocost, we encountered some list add/del corruptions of inner_walk list in ioc_timer_fn. The reason can be described as follows: cpu 0 cpu 1 ioc_qos_write ioc_qos_write ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ... rq_qos_add(q, rqos); } ... rq_qos_add(q, rqos); ... } When the io.cost.qos file is written by two cpus concurrently, rq_qos may be added to one disk twice. In that case, there will be two iocs enabled and running on one disk. They own different iocgs on their active list. In the ioc_timer_fn function, because of the iocgs from two iocs have the same root iocg, the root iocg's walk_list may be overwritten by each other and this leads to list add/del corruptions in building or destroying the inner_walk list. And so far, the blk-rq-qos framework works in case that one instance for one type rq_qos per queue by default. This patch make this explicit and also fix the crash above. Signed-off-by: Jinke Han <hanjinke.666@bytedance.com> Reviewed-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20220720093616.70584-1-hanjinke.666@bytedance.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-20 17:36:16 +08:00
blk-cgroup: pass a gendisk to blkcg_{de,}activate_policy Prepare for storing the blkcg information in the gendisk instead of the request_queue. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-17-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:57 +01:00
block: don't allow the same type rq_qos add more than once In our test of iocost, we encountered some list add/del corruptions of inner_walk list in ioc_timer_fn. The reason can be described as follows: cpu 0 cpu 1 ioc_qos_write ioc_qos_write ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ... rq_qos_add(q, rqos); } ... rq_qos_add(q, rqos); ... } When the io.cost.qos file is written by two cpus concurrently, rq_qos may be added to one disk twice. In that case, there will be two iocs enabled and running on one disk. They own different iocgs on their active list. In the ioc_timer_fn function, because of the iocgs from two iocs have the same root iocg, the root iocg's walk_list may be overwritten by each other and this leads to list add/del corruptions in building or destroying the inner_walk list. And so far, the blk-rq-qos framework works in case that one instance for one type rq_qos per queue by default. This patch make this explicit and also fix the crash above. Signed-off-by: Jinke Han <hanjinke.666@bytedance.com> Reviewed-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20220720093616.70584-1-hanjinke.666@bytedance.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-20 17:36:16 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: don't allow the same type rq_qos add more than once In our test of iocost, we encountered some list add/del corruptions of inner_walk list in ioc_timer_fn. The reason can be described as follows: cpu 0 cpu 1 ioc_qos_write ioc_qos_write ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ... rq_qos_add(q, rqos); } ... rq_qos_add(q, rqos); ... } When the io.cost.qos file is written by two cpus concurrently, rq_qos may be added to one disk twice. In that case, there will be two iocs enabled and running on one disk. They own different iocgs on their active list. In the ioc_timer_fn function, because of the iocgs from two iocs have the same root iocg, the root iocg's walk_list may be overwritten by each other and this leads to list add/del corruptions in building or destroying the inner_walk list. And so far, the blk-rq-qos framework works in case that one instance for one type rq_qos per queue by default. This patch make this explicit and also fix the crash above. Signed-off-by: Jinke Han <hanjinke.666@bytedance.com> Reviewed-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20220720093616.70584-1-hanjinke.666@bytedance.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-20 17:36:16 +08:00
blk-rq-qos: make rq_qos_add and rq_qos_del more useful Switch to passing a gendisk, and make rq_qos_add initialize all required fields and drop the not required q argument from rq_qos_del. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-14-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:54 +01:00
block: don't allow the same type rq_qos add more than once In our test of iocost, we encountered some list add/del corruptions of inner_walk list in ioc_timer_fn. The reason can be described as follows: cpu 0 cpu 1 ioc_qos_write ioc_qos_write ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ioc = q_to_ioc(queue); if (!ioc) { ioc = kzalloc(); ... rq_qos_add(q, rqos); } ... rq_qos_add(q, rqos); ... } When the io.cost.qos file is written by two cpus concurrently, rq_qos may be added to one disk twice. In that case, there will be two iocs enabled and running on one disk. They own different iocgs on their active list. In the ioc_timer_fn function, because of the iocgs from two iocs have the same root iocg, the root iocg's walk_list may be overwritten by each other and this leads to list add/del corruptions in building or destroying the inner_walk list. And so far, the blk-rq-qos framework works in case that one instance for one type rq_qos per queue by default. This patch make this explicit and also fix the crash above. Signed-off-by: Jinke Han <hanjinke.666@bytedance.com> Reviewed-by: Muchun Song <songmuchun@bytedance.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20220720093616.70584-1-hanjinke.666@bytedance.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-07-20 17:36:16 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: make iolatency avg_lat exponentially decay Currently, avg_lat is calculated by accumulating the mean of every window in a long running cumulative average. As time goes on, the metric becomes less and less useful due to the accumulated history. This patch reuses the same calculation done in load averages to make the avg_lat metric more lively. Unlike load averages, the avg only advances when a window elapses (due to an io). Idle periods extend the most recent window. Bucketing is used to limit the history of avg_lat by binding it to the window size. So, the window range for 1/exp (decay rate) is [1 min, 2.5 min) when windows elapse immediately. The current sample window size is exposed in the debug info to enable calculation of the window range. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-08-01 23:15:41 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-iolatency: clear use_delay when io.latency is set to zero If use_delay was non-zero when the latency target of a cgroup was set to zero, it will stay stuck until io.latency is enabled on the cgroup again. This keeps readahead disabled for the cgroup impacting performance negatively. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <jbacik@fb.com> Fixes: d70675121546 ("block: introduce blk-iolatency io controller") Cc: stable@vger.kernel.org # v4.19+ Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-13 15:30:37 -07:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
blk-iolatency: clear use_delay when io.latency is set to zero If use_delay was non-zero when the latency target of a cgroup was set to zero, it will stay stuck until io.latency is enabled on the cgroup again. This keeps readahead disabled for the cgroup impacting performance negatively. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <jbacik@fb.com> Fixes: d70675121546 ("block: introduce blk-iolatency io controller") Cc: stable@vger.kernel.org # v4.19+ Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-13 15:30:37 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: Restructure blkg_conf_prep() and friends We want to support lazy init of rq-qos policies so that iolatency is enabled lazily on configuration instead of gendisk initialization. The way blkg config helpers are structured now is a bit awkward for that. Let's restructure: * blkcg_conf_open_bdev() is renamed to blkg_conf_open_bdev(). The blkcg_ prefix was used because the bdev opening step is blkg-independent. However, the distinction is too subtle and confuses more than helps. Let's switch to blkg prefix so that it's consistent with the type and other helper names. * struct blkg_conf_ctx now remembers the original input string and is always initialized by the new blkg_conf_init(). * blkg_conf_open_bdev() is updated to take a pointer to blkg_conf_ctx like blkg_conf_prep() and can be called multiple times safely. Instead of modifying the double pointer to input string directly, blkg_conf_open_bdev() now sets blkg_conf_ctx->body. * blkg_conf_finish() is renamed to blkg_conf_exit() for symmetry and now must be called on all blkg_conf_ctx's which were initialized with blkg_conf_init(). Combined, this allows the users to either open the bdev first or do it altogether with blkg_conf_prep() which will help implementing lazy init of rq-qos policies. blkg_conf_init/exit() will also be used implement synchronization against device removal. This is necessary because iolat / iocost are configured through cgroupfs instead of one of the files under /sys/block/DEVICE. As cgroupfs operations aren't synchronized with block layer, the lazy init and other configuration operations may race against device removal. This patch makes blkg_conf_init/exit() used consistently for all cgroup-orginating configurations making them a good place to implement explicit synchronization. Users are updated accordingly. No behavior change is intended by this patch. v2: bfq wasn't updated in v1 causing a build error. Fixed. v3: Update the description to include future use of blkg_conf_init/exit() as synchronization points. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Yu Kuai <yukuai1@huaweicloud.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230413000649.115785-3-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:47 -10:00
block: remove init_mutex and open-code blk_iolatency_try_init Commit a13696b83da4 ("blk-iolatency: Make initialization lazy") adds a mutex named "init_mutex" in blk_iolatency_try_init for the race condition of initializing RQ_QOS_LATENCY. Now a new lock has been add to struct request_queue by commit a13bd91be223 ("block/rq_qos: protect rq_qos apis with a new lock"). And it has been held in blkg_conf_open_bdev before calling blk_iolatency_init. So it's not necessary to keep init_mutex in blk_iolatency_try_init, just remove it. Since init_mutex has been removed, blk_iolatency_try_init can be open-coded back to iolatency_set_limit() like ioc_qos_write(). Signed-off-by: Li Lingfeng <lilingfeng3@huawei.com> Reviewed-by: Michal Koutný <mkoutny@suse.com> Link: https://lore.kernel.org/r/20230810035111.2236335-1-lilingfeng@huaweicloud.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-08-10 11:51:11 +08:00
block: fix bad lockdep annotation in blk-iolatency A previous commit added a lockdep annotation, but botched it. Use the right type. Fixes: 4eb44d10766a ("block: remove init_mutex and open-code blk_iolatency_try_init") Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-08-10 17:24:53 -06:00
block: remove init_mutex and open-code blk_iolatency_try_init Commit a13696b83da4 ("blk-iolatency: Make initialization lazy") adds a mutex named "init_mutex" in blk_iolatency_try_init for the race condition of initializing RQ_QOS_LATENCY. Now a new lock has been add to struct request_queue by commit a13bd91be223 ("block/rq_qos: protect rq_qos apis with a new lock"). And it has been held in blkg_conf_open_bdev before calling blk_iolatency_init. So it's not necessary to keep init_mutex in blk_iolatency_try_init, just remove it. Since init_mutex has been removed, blk_iolatency_try_init can be open-coded back to iolatency_set_limit() like ioc_qos_write(). Signed-off-by: Li Lingfeng <lilingfeng3@huawei.com> Reviewed-by: Michal Koutný <mkoutny@suse.com> Link: https://lore.kernel.org/r/20230810035111.2236335-1-lilingfeng@huaweicloud.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-08-10 11:51:11 +08:00
blk-iolatency: Make initialization lazy Other rq_qos policies such as wbt and iocost are lazy-initialized when they are configured for the first time for the device but iolatency is initialized unconditionally from blkcg_init_disk() during gendisk init. Lazy init is beneficial because rq_qos policies add runtime overhead when initialized as every IO has to walk all registered rq_qos callbacks. This patch switches iolatency to lazy initialization too so that it only registered its rq_qos policy when it is first configured. Note that there is a known race condition between blkcg config file writes and del_gendisk() and this patch makes iolatency susceptible to it by exposing the init path to race against the deletion path. However, that problem already exists in iocost and is being worked on. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20230413000649.115785-5-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:49 -10:00
blkcg: Restructure blkg_conf_prep() and friends We want to support lazy init of rq-qos policies so that iolatency is enabled lazily on configuration instead of gendisk initialization. The way blkg config helpers are structured now is a bit awkward for that. Let's restructure: * blkcg_conf_open_bdev() is renamed to blkg_conf_open_bdev(). The blkcg_ prefix was used because the bdev opening step is blkg-independent. However, the distinction is too subtle and confuses more than helps. Let's switch to blkg prefix so that it's consistent with the type and other helper names. * struct blkg_conf_ctx now remembers the original input string and is always initialized by the new blkg_conf_init(). * blkg_conf_open_bdev() is updated to take a pointer to blkg_conf_ctx like blkg_conf_prep() and can be called multiple times safely. Instead of modifying the double pointer to input string directly, blkg_conf_open_bdev() now sets blkg_conf_ctx->body. * blkg_conf_finish() is renamed to blkg_conf_exit() for symmetry and now must be called on all blkg_conf_ctx's which were initialized with blkg_conf_init(). Combined, this allows the users to either open the bdev first or do it altogether with blkg_conf_prep() which will help implementing lazy init of rq-qos policies. blkg_conf_init/exit() will also be used implement synchronization against device removal. This is necessary because iolat / iocost are configured through cgroupfs instead of one of the files under /sys/block/DEVICE. As cgroupfs operations aren't synchronized with block layer, the lazy init and other configuration operations may race against device removal. This patch makes blkg_conf_init/exit() used consistently for all cgroup-orginating configurations making them a good place to implement explicit synchronization. Users are updated accordingly. No behavior change is intended by this patch. v2: bfq wasn't updated in v1 causing a build error. Fixed. v3: Update the description to include future use of blkg_conf_init/exit() as synchronization points. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Yu Kuai <yukuai1@huaweicloud.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230413000649.115785-3-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:47 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: Restructure blkg_conf_prep() and friends We want to support lazy init of rq-qos policies so that iolatency is enabled lazily on configuration instead of gendisk initialization. The way blkg config helpers are structured now is a bit awkward for that. Let's restructure: * blkcg_conf_open_bdev() is renamed to blkg_conf_open_bdev(). The blkcg_ prefix was used because the bdev opening step is blkg-independent. However, the distinction is too subtle and confuses more than helps. Let's switch to blkg prefix so that it's consistent with the type and other helper names. * struct blkg_conf_ctx now remembers the original input string and is always initialized by the new blkg_conf_init(). * blkg_conf_open_bdev() is updated to take a pointer to blkg_conf_ctx like blkg_conf_prep() and can be called multiple times safely. Instead of modifying the double pointer to input string directly, blkg_conf_open_bdev() now sets blkg_conf_ctx->body. * blkg_conf_finish() is renamed to blkg_conf_exit() for symmetry and now must be called on all blkg_conf_ctx's which were initialized with blkg_conf_init(). Combined, this allows the users to either open the bdev first or do it altogether with blkg_conf_prep() which will help implementing lazy init of rq-qos policies. blkg_conf_init/exit() will also be used implement synchronization against device removal. This is necessary because iolat / iocost are configured through cgroupfs instead of one of the files under /sys/block/DEVICE. As cgroupfs operations aren't synchronized with block layer, the lazy init and other configuration operations may race against device removal. This patch makes blkg_conf_init/exit() used consistently for all cgroup-orginating configurations making them a good place to implement explicit synchronization. Users are updated accordingly. No behavior change is intended by this patch. v2: bfq wasn't updated in v1 causing a build error. Fixed. v3: Update the description to include future use of blkg_conf_init/exit() as synchronization points. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Yu Kuai <yukuai1@huaweicloud.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230413000649.115785-3-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:47 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: Restructure blkg_conf_prep() and friends We want to support lazy init of rq-qos policies so that iolatency is enabled lazily on configuration instead of gendisk initialization. The way blkg config helpers are structured now is a bit awkward for that. Let's restructure: * blkcg_conf_open_bdev() is renamed to blkg_conf_open_bdev(). The blkcg_ prefix was used because the bdev opening step is blkg-independent. However, the distinction is too subtle and confuses more than helps. Let's switch to blkg prefix so that it's consistent with the type and other helper names. * struct blkg_conf_ctx now remembers the original input string and is always initialized by the new blkg_conf_init(). * blkg_conf_open_bdev() is updated to take a pointer to blkg_conf_ctx like blkg_conf_prep() and can be called multiple times safely. Instead of modifying the double pointer to input string directly, blkg_conf_open_bdev() now sets blkg_conf_ctx->body. * blkg_conf_finish() is renamed to blkg_conf_exit() for symmetry and now must be called on all blkg_conf_ctx's which were initialized with blkg_conf_init(). Combined, this allows the users to either open the bdev first or do it altogether with blkg_conf_prep() which will help implementing lazy init of rq-qos policies. blkg_conf_init/exit() will also be used implement synchronization against device removal. This is necessary because iolat / iocost are configured through cgroupfs instead of one of the files under /sys/block/DEVICE. As cgroupfs operations aren't synchronized with block layer, the lazy init and other configuration operations may race against device removal. This patch makes blkg_conf_init/exit() used consistently for all cgroup-orginating configurations making them a good place to implement explicit synchronization. Users are updated accordingly. No behavior change is intended by this patch. v2: bfq wasn't updated in v1 causing a build error. Fixed. v3: Update the description to include future use of blkg_conf_init/exit() as synchronization points. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Yu Kuai <yukuai1@huaweicloud.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230413000649.115785-3-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:47 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: iolatency: avoid 64-bit division On 32-bit architectures, dividing a 64-bit number needs to use the do_div() function or something like it to avoid a link failure: block/blk-iolatency.o: In function `iolatency_prfill_limit': blk-iolatency.c:(.text+0x8cc): undefined reference to `__aeabi_uldivmod' Using div_u64() gives us the best output and avoids the need for an explicit cast. Fixes: d70675121546 ("block: introduce blk-iolatency io controller") Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-10 17:21:34 +02:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-cgroup: always terminate io.stat lines With the removal of seq_get_buf in blkcg_print_one_stat, we cannot make adding the newline conditional on there being relevant stats because the name was already written out unconditionally. Otherwise we may end up with multiple device names in one line which is confusing and doesn't follow the nested-keyed file format. Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com> Fixes: 252c651a4c85 ("blk-cgroup: stop using seq_get_buf") Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20220111083159.42340-1-w.bumiller@proxmox.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-01-11 09:31:59 +01:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
blk-cgroup: stop using seq_get_buf seq_get_buf is a crutch that undoes all the memory safety of the seq_file interface. Use the normal seq_printf interfaces instead. Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20210810152623.1796144-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2021-08-10 17:26:23 +02:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
blk-cgroup: always terminate io.stat lines With the removal of seq_get_buf in blkcg_print_one_stat, we cannot make adding the newline conditional on there being relevant stats because the name was already written out unconditionally. Otherwise we may end up with multiple device names in one line which is confusing and doesn't follow the nested-keyed file format. Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com> Fixes: 252c651a4c85 ("blk-cgroup: stop using seq_get_buf") Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20220111083159.42340-1-w.bumiller@proxmox.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-01-11 09:31:59 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blkcg: allow blkcg_policy->pd_stat() to print non-debug info too Currently, ->pd_stat() is called only when moduleparam blkcg_debug_stats is set which prevents it from printing non-debug policy-specific statistics. Let's move debug testing down so that ->pd_stat() can print non-debug stat too. This patch doesn't cause any visible behavior change. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-07-16 07:58:31 -07:00
blk-cgroup: always terminate io.stat lines With the removal of seq_get_buf in blkcg_print_one_stat, we cannot make adding the newline conditional on there being relevant stats because the name was already written out unconditionally. Otherwise we may end up with multiple device names in one line which is confusing and doesn't follow the nested-keyed file format. Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com> Fixes: 252c651a4c85 ("blk-cgroup: stop using seq_get_buf") Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20220111083159.42340-1-w.bumiller@proxmox.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-01-11 09:31:59 +01:00
blkcg: allow blkcg_policy->pd_stat() to print non-debug info too Currently, ->pd_stat() is called only when moduleparam blkcg_debug_stats is set which prevents it from printing non-debug policy-specific statistics. Let's move debug testing down so that ->pd_stat() can print non-debug stat too. This patch doesn't cause any visible behavior change. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-07-16 07:58:31 -07:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
blk-cgroup: stop using seq_get_buf seq_get_buf is a crutch that undoes all the memory safety of the seq_file interface. Use the normal seq_printf interfaces instead. Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20210810152623.1796144-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2021-08-10 17:26:23 +02:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
blk-cgroup: stop using seq_get_buf seq_get_buf is a crutch that undoes all the memory safety of the seq_file interface. Use the normal seq_printf interfaces instead. Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20210810152623.1796144-2-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2021-08-10 17:26:23 +02:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-cgroup: pass a gendisk to pd_alloc_fn No need to the request_queue here, pass a gendisk and extract the node ids from that. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-18-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:58 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-cgroup: pass a gendisk to pd_alloc_fn No need to the request_queue here, pass a gendisk and extract the node ids from that. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Andreas Herrmann <aherrmann@suse.de> Acked-by: Tejun Heo <tj@kernel.org> Link: https://lore.kernel.org/r/20230203150400.3199230-18-hch@lst.de Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-02-03 16:03:58 +01:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: s/blkcg_rq_qos/iolat_rq_qos/ The name was too generic given that there are multiple blkcg rq-qos policies. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20230413000649.115785-4-tj@kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2023-04-12 14:06:48 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: add blk_time_get_ns() and blk_time_get() helpers Convert any user of ktime_get_ns() to use blk_time_get_ns(), and ktime_get() to blk_time_get(), so we have a unified API for querying the current time in nanoseconds or as ktime. No functional changes intended, this patch just wraps ktime_get_ns() and ktime_get() with a block helper. Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2024-01-15 14:45:07 -07:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: introduce blk_queue_rot() To check if a request queue is for a rotational device, a double negation is needed with the pattern "!blk_queue_nonrot(q)". Simplify this with the introduction of the helper blk_queue_rot() which tests if a requests queue limit has the BLK_FEAT_ROTATIONAL feature set. All call sites of blk_queue_nonrot() are modified to use blk_queue_rot() and blk_queue_nonrot() definition removed. No functional changes. Signed-off-by: Damien Le Moal <dlemoal@kernel.org> Reviewed-by: Nitesh Shetty <nj.shetty@samsung.com> Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2026-01-29 16:27:15 +09:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: use a percentile approache for ssd's We use an average latency approach for determining if we're missing our latency target. This works well for rotational storage where we have generally consistent latencies, but for ssd's and other low latency devices you have more of a spikey behavior, which means we often won't throttle misbehaving groups because a lot of IO completes at drastically faster times than our latency target. Instead keep track of how many IO's miss our target and how many IO's are done in our time window. If the p(90) latency is above our target then we know we need to throttle. With this change in place we are seeing the same throttling behavior with our testcase on ssd's as we see with rotational drives. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:42 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: keep track of previous windows stats We apply a smoothing to the scale changes in order to keep sawtoothy behavior from occurring. However our window for checking if we've missed our target can sometimes be lower than the smoothing interval (500ms), especially on faster drives like ssd's. In order to deal with this keep track of the running tally of the previous intervals that we threw away because we had already done a scale event recently. This is needed for the ssd case as these low latency drives will have bursts of latency, and if it happens to be ok for the window that directly follows the opening of the scale window we could unthrottle when previous windows we were missing our target. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-09-28 13:45:43 -04:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Replace struct rq_depth with unsigned int in struct iolatency_grp We only need a max queue depth for every iolatency to limit the inflight io number. Replace struct rq_depth with unsigned int to simplfy "struct iolatency_grp" and save memory. Signed-off-by: Kemeng Shi <shikemeng@huawei.com> Reviewed-by: Josef Bacik <josef@toxicpanda.com> Link: https://lore.kernel.org/r/20221018111240.22612-4-shikemeng@huawei.com Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-10-18 19:12:40 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
blk-iolatency: Fix inflight count imbalances and IO hangs on offline iolatency needs to track the number of inflight IOs per cgroup. As this tracking can be expensive, it is disabled when no cgroup has iolatency configured for the device. To ensure that the inflight counters stay balanced, iolatency_set_limit() freezes the request_queue while manipulating the enabled counter, which ensures that no IO is in flight and thus all counters are zero. Unfortunately, iolatency_set_limit() isn't the only place where the enabled counter is manipulated. iolatency_pd_offline() can also dec the counter and trigger disabling. As this disabling happens without freezing the q, this can easily happen while some IOs are in flight and thus leak the counts. This can be easily demonstrated by turning on iolatency on an one empty cgroup while IOs are in flight in other cgroups and then removing the cgroup. Note that iolatency shouldn't have been enabled elsewhere in the system to ensure that removing the cgroup disables iolatency for the whole device. The following keeps flipping on and off iolatency on sda: echo +io > /sys/fs/cgroup/cgroup.subtree_control while true; do mkdir -p /sys/fs/cgroup/test echo '8:0 target=100000' > /sys/fs/cgroup/test/io.latency sleep 1 rmdir /sys/fs/cgroup/test sleep 1 done and there's concurrent fio generating direct rand reads: fio --name test --filename=/dev/sda --direct=1 --rw=randread \ --runtime=600 --time_based --iodepth=256 --numjobs=4 --bs=4k while monitoring with the following drgn script: while True: for css in css_for_each_descendant_pre(prog['blkcg_root'].css.address_of_()): for pos in hlist_for_each(container_of(css, 'struct blkcg', 'css').blkg_list): blkg = container_of(pos, 'struct blkcg_gq', 'blkcg_node') pd = blkg.pd[prog['blkcg_policy_iolatency'].plid] if pd.value_() == 0: continue iolat = container_of(pd, 'struct iolatency_grp', 'pd') inflight = iolat.rq_wait.inflight.counter.value_() if inflight: print(f'inflight={inflight} {disk_name(blkg.q.disk).decode("utf-8")} ' f'{cgroup_path(css.cgroup).decode("utf-8")}') time.sleep(1) The monitoring output looks like the following: inflight=1 sda /user.slice inflight=1 sda /user.slice ... inflight=14 sda /user.slice inflight=13 sda /user.slice inflight=17 sda /user.slice inflight=15 sda /user.slice inflight=18 sda /user.slice inflight=17 sda /user.slice inflight=20 sda /user.slice inflight=19 sda /user.slice <- fio stopped, inflight stuck at 19 inflight=19 sda /user.slice inflight=19 sda /user.slice If a cgroup with stuck inflight ends up getting throttled, the throttled IOs will never get issued as there's no completion event to wake it up leading to an indefinite hang. This patch fixes the bug by unifying enable handling into a work item which is automatically kicked off from iolatency_set_min_lat_nsec() which is called from both iolatency_set_limit() and iolatency_pd_offline() paths. Punting to a work item is necessary as iolatency_pd_offline() is called under spinlocks while freezing a request_queue requires a sleepable context. This also simplifies the code reducing LOC sans the comments and avoids the unnecessary freezes which were happening whenever a cgroup's latency target is newly set or cleared. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Josef Bacik <josef@toxicpanda.com> Cc: Liu Bo <bo.liu@linux.alibaba.com> Fixes: 8c772a9bfc7c ("blk-iolatency: fix IO hang due to negative inflight counter") Cc: stable@vger.kernel.org # v5.0+ Link: https://lore.kernel.org/r/Yn9ScX6Nx2qIiQQi@slm.duckdns.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2022-05-13 20:55:45 -10:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00
block: Remove redundant 'return' statement Remove redundant 'return' statement for 'void' functions. Signed-off-by: Baolin Wang <baolin.wang@linux.alibaba.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-09-28 08:42:26 +08:00
block: introduce blk-iolatency io controller Current IO controllers for the block layer are less than ideal for our use case. The io.max controller is great at hard limiting, but it is not work conserving. This patch introduces io.latency. You provide a latency target for your group and we monitor the io in short windows to make sure we are not exceeding those latency targets. This makes use of the rq-qos infrastructure and works much like the wbt stuff. There are a few differences from wbt - It's bio based, so the latency covers the whole block layer in addition to the actual io. - We will throttle all IO types that comes in here if we need to. - We use the mean latency over the 100ms window. This is because writes can be particularly fast, which could give us a false sense of the impact of other workloads on our protected workload. - By default there's no throttling, we set the queue_depth to INT_MAX so that we can have as many outstanding bio's as we're allowed to. Only at throttle time do we pay attention to the actual queue depth. - We backcharge cgroups for root cg issued IO and induce artificial delays in order to deal with cases like metadata only or swap heavy workloads. In testing this has worked out relatively well. Protected workloads will throttle noisy workloads down to 1 io at time if they are doing normal IO on their own, or induce up to a 1 second delay per syscall if they are doing a lot of root issued IO (metadata/swap IO). Our testing has revolved mostly around our production web servers where we have hhvm (the web server application) in a protected group and everything else in another group. We see slightly higher requests per second (RPS) on the test tier vs the control tier, and much more stable RPS across all machines in the test tier vs the control tier. Another test we run is a slow memory allocator in the unprotected group. Before this would eventually push us into swap and cause the whole box to die and not recover at all. With these patches we see slight RPS drops (usually 10-15%) before the memory consumer is properly killed and things recover within seconds. Signed-off-by: Josef Bacik <jbacik@fb.com> Acked-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-07-03 11:15:01 -04:00