diff options
| author | Martin Schwenke <martin@meltin.net> | 2012-05-16 13:29:58 +1000 |
|---|---|---|
| committer | Martin Schwenke <martin@meltin.net> | 2012-05-16 17:19:55 +1000 |
| commit | ffbe59bd44fee532e0167b56f4f031dc3d673380 (patch) | |
| tree | 96751dc06832deb4061a54039dfd0b1aebfde9fb | |
| parent | 92eb004162ac1729a3ca5210a0c97257e86cd0c6 (diff) | |
Eventscripts: restart lockd in the background when going unhealthy
Sometimes the restart can hang when there are I/O problems. Then the
eventscript times out and gets killed so the node never marked as
unhealthy.
Restarting in the background avoids this.
Signed-off-by: Martin Schwenke <martin@meltin.net>
(This used to be ctdb commit 13acd58c41fba1a33894fbd654fed69ea0eac322)
| -rwxr-xr-x | ctdb/config/events.d/60.nfs | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/ctdb/config/events.d/60.nfs b/ctdb/config/events.d/60.nfs index e8ac61fe43..6ccec9f245 100755 --- a/ctdb/config/events.d/60.nfs +++ b/ctdb/config/events.d/60.nfs @@ -98,7 +98,7 @@ case "$1" in # check that lockd responds to rpc requests nfs_check_rpc_service "lockd" \ - -ge 15 "verbose restart unhealthy" \ + -ge 15 "verbose restart:b unhealthy" \ -eq 10 "restart:bs" # mountd is sometimes not started correctly on RHEL5 |