Fix hung z_zvol tasks during 'zfs receive'
During a receive operation zvol_create_minors_impl() can wait needlessly for the prefetch thread because both share the same tasks queue. This results in hung tasks: <3>INFO: task z_zvol:5541 blocked for more than 120 seconds. <3> Tainted: P O 3.16.0-4-amd64 <3>"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. The first z_zvol:5541 (zvol_task_cb) is waiting for the long running traverse_prefetch_thread:260 root@linux:~# cat /proc/spl/taskq taskq act nthr spwn maxt pri mina spl_system_taskq/0 1 2 0 64 100 1 active: [260]traverse_prefetch_thread [zfs](0xffff88003347ae40) wait: 5541 spl_delay_taskq/0 0 1 0 4 100 1 delay: spa_deadman [zfs](0xffff880039924000) z_zvol/1 1 1 0 1 120 1 active: [5541]zvol_task_cb [zfs](0xffff88001fde6400) pend: zvol_task_cb [zfs](0xffff88001fde6800) This change adds a dedicated, per-pool, prefetch taskq to prevent the traverse code from monopolizing the global (and limited) system_taskq by inappropriately scheduling long running tasks on it. Reviewed-by: Albert Lee <trisk@forkgnu.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: loli10K <ezomori.nozomu@gmail.com> Closes #6330 Closes #6890 Closes #7343
This commit is contained in:
parent
3b118f0a34
commit
fd01167ffd
|
@ -275,6 +275,7 @@ struct spa {
|
|||
spa_stats_t spa_stats; /* assorted spa statistics */
|
||||
hrtime_t spa_ccw_fail_time; /* Conf cache write fail time */
|
||||
taskq_t *spa_zvol_taskq; /* Taskq for minor management */
|
||||
taskq_t *spa_prefetch_taskq; /* Taskq for prefetch threads */
|
||||
uint64_t spa_multihost; /* multihost aware (mmp) */
|
||||
mmp_thread_t spa_mmp; /* multihost mmp thread */
|
||||
|
||||
|
|
|
@ -31,6 +31,7 @@
|
|||
#include <sys/dsl_pool.h>
|
||||
#include <sys/dnode.h>
|
||||
#include <sys/spa.h>
|
||||
#include <sys/spa_impl.h>
|
||||
#include <sys/zio.h>
|
||||
#include <sys/dmu_impl.h>
|
||||
#include <sys/sa.h>
|
||||
|
@ -623,7 +624,7 @@ traverse_impl(spa_t *spa, dsl_dataset_t *ds, uint64_t objset, blkptr_t *rootbp,
|
|||
}
|
||||
|
||||
if (!(flags & TRAVERSE_PREFETCH_DATA) ||
|
||||
taskq_dispatch(system_taskq, traverse_prefetch_thread,
|
||||
taskq_dispatch(spa->spa_prefetch_taskq, traverse_prefetch_thread,
|
||||
td, TQ_NOQUEUE) == TASKQID_INVALID)
|
||||
pd->pd_exited = B_TRUE;
|
||||
|
||||
|
|
|
@ -1182,6 +1182,14 @@ spa_activate(spa_t *spa, int mode)
|
|||
spa->spa_zvol_taskq = taskq_create("z_zvol", 1, defclsyspri,
|
||||
1, INT_MAX, 0);
|
||||
|
||||
/*
|
||||
* Taskq dedicated to prefetcher threads: this is used to prevent the
|
||||
* pool traverse code from monopolizing the global (and limited)
|
||||
* system_taskq by inappropriately scheduling long running tasks on it.
|
||||
*/
|
||||
spa->spa_prefetch_taskq = taskq_create("z_prefetch", boot_ncpus,
|
||||
defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC);
|
||||
|
||||
/*
|
||||
* The taskq to upgrade datasets in this pool. Currently used by
|
||||
* feature SPA_FEATURE_USEROBJ_ACCOUNTING.
|
||||
|
@ -1211,6 +1219,11 @@ spa_deactivate(spa_t *spa)
|
|||
spa->spa_zvol_taskq = NULL;
|
||||
}
|
||||
|
||||
if (spa->spa_prefetch_taskq) {
|
||||
taskq_destroy(spa->spa_prefetch_taskq);
|
||||
spa->spa_prefetch_taskq = NULL;
|
||||
}
|
||||
|
||||
if (spa->spa_upgrade_taskq) {
|
||||
taskq_destroy(spa->spa_upgrade_taskq);
|
||||
spa->spa_upgrade_taskq = NULL;
|
||||
|
|
Loading…
Reference in New Issue