60.nfs

Try to restart LOCKD after 10 failures and
flag the node as unhealthy after 15 failures

diff --git a/config/events.d/60.nfs b/config/events.d/60.nfs
index a8fe24349618af0b7d4abad927a3c972df8ba26c..87955df92ccd4595210943f734d83a1f0afe657c 100755 (executable)
--- a/config/events.d/60.nfs
+++ b/config/events.d/60.nfs
@@ -107,12 +107,29 @@ case "$1" in
        } || exit $?
 
        # check that lockd responds to rpc requests
-       ctdb_check_rpc "LOCKD" 100021 1 || {
-               echo "Trying to restart lock manager service"
-               startstop_nfs restart
-               startstop_nfslock restart
-               exit 1
-       }
+       if ctdb_check_rpc "LOCKD" 100021 1 >/dev/null ; then
+               (service_name="lockd"; ctdb_counter_init)
+       else
+               (
+                       service_name="lockd"
+                       ctdb_counter_incr
+
+                       ctdb_check_counter_equal 10 || {
+                               echo "Trying to restart NFS lock service"
+                               startstop_nfs restart >/dev/null 2>&1 &
+                               startstop_nfslock restart  >/dev/null 2>&1 &
+                               exit 0
+                       }
+
+                       ctdb_check_counter_limit 15 quiet >/dev/null
+       ) || {
+                       echo "$ctdb_check_rpc_out"
+                       echo "Trying to restart NFS lock service"
+                       startstop_nfs restart
+                       startstop_nfslock restart
+                       exit 1
+               }
+       fi
 
        # mount needs special handling since it is sometimes not started
        # correctly on RHEL5