We need to change the capacity in all nodes after one node
finishs reshape. And as we did before, we can't change the
capacity directly in md_do_sync, instead, the capacity should
be only changed in update_size or received CHANGE_CAPACITY
msg.
So master node calls update_size after completes reshape in
md_reap_sync_thread, but we need to skip ops->update_size if
MD_CLOSING is set since reshaping could not be finish.
Reviewed-by: NeilBrown <neilb@xxxxxxxx>
Signed-off-by: Guoqing Jiang <gqjiang@xxxxxxxx>
---
drivers/md/md.c | 21 ++++++++++++++++++---
1 file changed, 18 insertions(+), 3 deletions(-)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 43b6d2527a66..77d1afb4d480 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -8627,8 +8627,10 @@ void md_do_sync(struct md_thread *thread)
mddev_lock_nointr(mddev);
md_set_array_sectors(mddev, mddev->pers->size(mddev, 0, 0));
mddev_unlock(mddev);
- set_capacity(mddev->gendisk, mddev->array_sectors);
- revalidate_disk(mddev->gendisk);
+ if (!mddev_is_clustered(mddev)) {
+ set_capacity(mddev->gendisk, mddev->array_sectors);
+ revalidate_disk(mddev->gendisk);
+ }
}
spin_lock(&mddev->lock);
@@ -8972,6 +8974,8 @@ EXPORT_SYMBOL(md_check_recovery);
void md_reap_sync_thread(struct mddev *mddev)
{
struct md_rdev *rdev;
+ sector_t old_dev_sectors = mddev->dev_sectors;
+ bool is_reshaped = false;
/* resync has finished, collect result */
md_unregister_thread(&mddev->sync_thread);
@@ -8986,8 +8990,11 @@ void md_reap_sync_thread(struct mddev *mddev)
}
}
if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
- mddev->pers->finish_reshape)
+ mddev->pers->finish_reshape) {
mddev->pers->finish_reshape(mddev);
+ if (mddev_is_clustered(mddev))
+ is_reshaped = true;
+ }
/* If array is no-longer degraded, then any saved_raid_disk
* information must be scrapped.
@@ -9008,6 +9015,14 @@ void md_reap_sync_thread(struct mddev *mddev)
clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
+ /*
+ * We call md_cluster_ops->update_size here because sync_size could
+ * be changed by md_update_sb, and MD_RECOVERY_RESHAPE is cleared,
+ * so it is time to update size across cluster.
+ */
+ if (mddev_is_clustered(mddev) && is_reshaped
+ && !test_bit(MD_CLOSING, &mddev->flags))
+ md_cluster_ops->update_size(mddev, old_dev_sectors);
wake_up(&resync_wait);
/* flag recovery needed just to double check */
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
--
2.12.3
