LU-8910 osp: Add correct handling of errors to osp_statfs_interpret

MDT's statfs info could be disagreed with OST's info for a very long time.
If osp_statfs_update() is called and extends the timeout 1000*obd_timeout
into the future but then osp_statfs_interpret() hits an error it
will never reset the timeout.
Now when osp_update_statfs request fails osp_statfs_interpret causes
osp_precreate_cleanup_orphans to send new one after 10 seconds.

Change-Id: Ib282d806ba4932db5c72df34905988f96de99297
Cray-bug-id: MRP-3892
Signed-off-by: Sergey Cheremencev <c17829@cray.com>
Reviewed-on: https://review.whamcloud.com/24167
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Tested-by: Jenkins
Reviewed-by: Alexandr Boyko <c17825@cray.com>
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>

diff --git a/lustre/osp/osp_precreate.c b/lustre/osp/osp_precreate.c
index 97ed157..8fbdcb4 100644 (file)
--- a/lustre/osp/osp_precreate.c
+++ b/lustre/osp/osp_precreate.c
@@ -150,7 +150,9 @@ static int osp_statfs_interpret(const struct lu_env *env,
 
        RETURN(0);
 out:
-       /* couldn't update statfs, try again as soon as possible */
+       /* couldn't update statfs, try again with a small delay */
+       d->opd_statfs_fresh_till = ktime_add_ns(ktime_get(), 10 * NSEC_PER_SEC);
+       d->opd_statfs_update_in_progress = 0;
        if (d->opd_pre != NULL && osp_precreate_running(d))
                wake_up(&d->opd_pre_waitq);