Fix "Detach spare vdev in case if resilvering does not happen"

Spare vdev should detach from the pool when a disk is reinserted.
However, spare detachment depends on the completion of resilvering,
and if resilver does not schedule, the spare vdev keeps attached to
the pool until the next resilvering. When a zfs pool contains
several disks (25+ mirror), resilvering does not always happen when
a disk is reinserted. In this patch, spare vdev is manually detached
from the pool when resilvering does not occur and it has been tested
on both Linux and FreeBSD.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Ameer Hamza <ahamza@ixsystems.com>
Closes #14722

diff --git a/include/sys/spa.h b/include/sys/spa.h
index 79c46aa07709ae0dd11ce8bea3ab22e12c0ac29e..a1c102020223eeb70e72fe6d770feb67e7cff392 100644 (file)
--- a/include/sys/spa.h
+++ b/include/sys/spa.h
@@ -787,6 +787,7 @@ extern int bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx);
 #define        SPA_ASYNC_L2CACHE_REBUILD               0x800
 #define        SPA_ASYNC_L2CACHE_TRIM                  0x1000
 #define        SPA_ASYNC_REBUILD_DONE                  0x2000
+#define        SPA_ASYNC_DETACH_SPARE                  0x4000
 
 /* device manipulation */
 extern int spa_vdev_add(spa_t *spa, nvlist_t *nvroot);

diff --git a/module/zfs/spa.c b/module/zfs/spa.c
index dc202978c0f6629b62777985dc48f6a18ad491b1..e76700a9caa3c7ed104393ddf0610d951cc50849 100644 (file)
--- a/module/zfs/spa.c
+++ b/module/zfs/spa.c
@@ -7058,7 +7058,7 @@ spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing,
  * Detach a device from a mirror or replacing vdev.
  *
  * If 'replace_done' is specified, only detach if the parent
- * is a replacing vdev.
+ * is a replacing or a spare vdev.
  */
 int
 spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done)
@@ -8281,7 +8281,8 @@ spa_async_thread(void *arg)
         * If any devices are done replacing, detach them.
         */
        if (tasks & SPA_ASYNC_RESILVER_DONE ||
-           tasks & SPA_ASYNC_REBUILD_DONE) {
+           tasks & SPA_ASYNC_REBUILD_DONE ||
+           tasks & SPA_ASYNC_DETACH_SPARE) {
                spa_vdev_resilver_done(spa);
        }
 

diff --git a/module/zfs/vdev.c b/module/zfs/vdev.c
index 7cf858c05051f18ceb1ad465e1604cc2fdbf79ef..241be8fd856c23e9bb67f3687e715d6863269ea6 100644 (file)
--- a/module/zfs/vdev.c
+++ b/module/zfs/vdev.c
@@ -4183,9 +4183,19 @@ vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate)
 
        if (wasoffline ||
            (oldstate < VDEV_STATE_DEGRADED &&
-           vd->vdev_state >= VDEV_STATE_DEGRADED))
+           vd->vdev_state >= VDEV_STATE_DEGRADED)) {
                spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_ONLINE);
 
+               /*
+                * Asynchronously detach spare vdev if resilver or
+                * rebuild is not required
+                */
+               if (vd->vdev_unspare &&
+                   !dsl_scan_resilvering(spa->spa_dsl_pool) &&
+                   !dsl_scan_resilver_scheduled(spa->spa_dsl_pool) &&
+                   !vdev_rebuild_active(tvd))
+                       spa_async_request(spa, SPA_ASYNC_DETACH_SPARE);
+       }
        return (spa_vdev_state_exit(spa, vd, 0));
 }