Revert "try to restart statd everytime it fails, not just the first time"

[ctdb.git] / config / events.d / 60.nfs

diff --git a/config/events.d/60.nfs b/config/events.d/60.nfs
index 0d59c7a7004af2283ec3743b7991d06d6b20a259..b5cd81921320591c6e27137f2fa0eebb32684d78 100755 (executable)
--- a/config/events.d/60.nfs
+++ b/config/events.d/60.nfs
@@ -21,6 +21,8 @@ case $cmd in
        /bin/mkdir -p $CTDB_BASE/state/statd/ip
        /bin/mkdir -p $STATD_SHARED_DIRECTORY
 
+       /bin/rm -f $CTDB_BASE/state/statd/statd.restart >/dev/null 2>/dev/null
+
        # make sure nfs is stopped before we start it, or it may get a bind error
        startstop_nfs stop
        startstop_nfs start
@@ -67,15 +69,29 @@ case $cmd in
 
       monitor)
        # check that statd responds to rpc requests
-       # if statd is not running we try to restart it
-       rpcinfo -u localhost 100024 1 > /dev/null || {
-               RPCSTATDOPTS=""
-               [ -n "$STATD_HOSTNAME" ] && RPCSTATDOPTS="$RPCSTATDOPTS -n $STATD_HOSTNAME"
-               [ -n "$STATD_PORT" ] && RPCSTATDOPTS="$RPCSTATDOPTS -p $STATD_PORT"
-               [ -n "$STATD_OUTGOING_PORT" ] && RPCSTATDOPTS="$RPCSTATDOPTS -o $STATD_OUTGOING_PORT"
-               rpc.statd $RPCSTATDOPTS 
-               echo "ERROR: STATD is not responding. Trying to restart it. [rpc.statd $RPCSTATDOPTS]"
-       }
+       # if statd is not running we try to restart it once and wait
+       # for the next monitoring event to verify if it is running or not
+       # if it still fails we fail and mark the node as UNHEALTHY
+       if [ -f $CTDB_BASE/state/statd/statd.restart ]; then
+               # statd was restarted, see if it came up ok
+               rpcinfo -u localhost 100024 1 > /dev/null || {
+                       echo "ERROR: Failed to restart STATD"
+                       exit 1
+               }
+               echo "STATD successfully restarted."
+               /bin/rm -f $CTDB_BASE/state/statd/statd.restart
+       else
+               rpcinfo -u localhost 100024 1 > /dev/null || {
+                       RPCSTATDOPTS=""
+                       [ -n "$STATD_HOSTNAME" ] && RPCSTATDOPTS="$RPCSTATDOPTS -n $STATD_HOSTNAME"
+                       [ -n "$STATD_PORT" ] && RPCSTATDOPTS="$RPCSTATDOPTS -p $STATD_PORT"
+                       [ -n "$STATD_OUTGOING_PORT" ] && RPCSTATDOPTS="$RPCSTATDOPTS -o $STATD_OUTGOING_PORT"
+                       rpc.statd $RPCSTATDOPTS 
+                       echo "ERROR: STATD is not responding. Trying to restart it. [rpc.statd $RPCSTATDOPTS]"
+                       touch $CTDB_BASE/state/statd/statd.restart
+               }
+       fi
+
 
 
        # check that NFS responds to rpc requests