mddev->in_sync is always zero in clustered array. It will trigger
raid1 report misleading message on assembling:
> md/raid1:md0: not clean -- starting background reconstruction
> md/raid1:md0: active with 2 out of 2 mirrors
This patch allows clustered array to set correct resync_offset.
If this patch is accepted, mdadm can remove commit f7a6246bab1541
("super1.c: avoid useless sync when bitmap switches from clustered
to none")
Signed-off-by: Heming Zhao <heming.zhao@xxxxxxxx>
---
drivers/md/md.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 21da0c48f6c2..a8654d7a1ed9 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -2010,7 +2010,7 @@ static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev)
sb->utime = cpu_to_le64((__u64)mddev->utime);
sb->events = cpu_to_le64(mddev->events);
- if (mddev->in_sync)
+ if (mddev->in_sync || mddev_is_clustered(mddev))
sb->resync_offset = cpu_to_le64(mddev->recovery_cp);
else if (test_bit(MD_JOURNAL_CLEAN, &mddev->flags))
sb->resync_offset = cpu_to_le64(MaxSector);
--
2.30.0
