common-ha: reliable grace using pacemaker notify actions

Using *-dead_ip-1 resources to track on which nodes the ganesha.nfsd
had died was found to be unreliable.

Running `pcs status` in the ganesha_grace monitor action was seen to
time out during failover; the HA devs opined that it was, generally,
not a good idea to run `pcs status` in a monitor action in any event.
They suggested using the notify feature, where the resources on all
the nodes are notified when a clone resource agent dies.

This change adds a notify action to the ganesha_grace RA. The ganesha_mon
RA monitors its ganesha.nfsd daemon. While the daemon is running, it
creates two attributes: ganesha-active and grace-active. When the daemon
stops for any reason, the attributes are deleted. Deleting the
ganesha-active attribute triggers the failover of the virtual IP (the
IPaddr RA) to another node where ganesha.nfsd is still running. The
ganesha_grace RA monitors the grace-active attribute. When the
grace-active attibute is deleted, the ganesha_grace RA stops, and will
not restart. This triggers pacemaker to trigger the notify action in
the ganesha_grace RAs on the other nodes in the cluster; which send a
DBUS message to their ganesha.nfsd.

(N.B. grace-active is a bit of a misnomer. while the grace-active
attribute exists, everything is normal and healthy. Deleting the
attribute triggers putting the surviving ganesha.nfsds into GRACE.)

To ensure that the remaining/surviving ganesha.nfsds are put into
NFS-GRACE before the IPaddr (virtual IP) fails over there is a short
delay (sleep) between deleting the grace-active attribute and the
ganesha-active attribute. To summarize:
  1. on node 2 ganesha_mon:monitor notices that ganesha.nfsd has died
  2. on node 2 ganesha_mon:monitor deletes its grace-active attribute
  3. on node 2 ganesha_grace:monitor notices that grace-active is gone
     and returns OCF_ERR_GENERIC, a.k.a. new error. When pacemaker
     tries to (re)start ganesha_grace, its start action will return
     OCF_NOT_RUNNING, a.k.a. known error, don't attempt further
     restarts.
  4. on nodes 1, 3, etc., ganesha_grace:notify receives a post-stop
     notification indicating that node 2 is gone, and sends a DBUS
     message to its ganesha.nfsd putting it into NFS-GRACE.
  5. on node 2 ganesha_mon:monitor waits a short period, then deletes
     its ganesha-active attribute. This triggers the IPaddr (virt IP)
     failover according to constraint location rules.

ganesha_nfsd modified to run for the duration, start action is invoked
to setup the /var/lib/nfs symlink, stop action is invoked to restore it.
ganesha-ha.sh modified accordingly to create it as a clone resource.

Change-Id: Iad60b0c5222bbd55ef95c8b8f955e791caa3ffd0
BUG: 1290865
Signed-off-by: Kaleb S KEITHLEY <kkeithle@redhat.com>
Reviewed-on: http://review.gluster.org/12964
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>

3 files changed, 199 insertions, 141 deletions

diff --git a/extras/ganesha/ocf/ganesha_grace b/extras/ganesha/ocf/ganesha_grace
index 75ec16c0fd1..a82c9af417a 100644
--- a/extras/ganesha/ocf/ganesha_grace
+++ b/extras/ganesha/ocf/ganesha_grace
@@ -36,6 +36,9 @@ else
 . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
 fi
 
+OCF_RESKEY_grace_active_default="grace-active"
+: ${OCF_RESKEY_grace_active=${OCF_RESKEY_grace_active_default}}
+
 ganesha_meta_data() {
         cat <<END
 <?xml version="1.0"?>
@@ -51,19 +54,25 @@ resource agent for nfs-ganesha.
 <shortdesc lang="en">Manages the user-space nfs-ganesha NFS server</shortdesc>
 
 <parameters>
+<parameter name="grace_active">
+<longdesc lang="en">NFS-Ganesha grace active attribute</longdesc>
+<shortdesc lang="en">NFS-Ganesha grace active attribute</shortdesc>
+<content type="string" default="grace-active" />
+</parameter>
 </parameters>
 
 <actions>
 <action name="start"   timeout="40s" />
 <action name="stop"    timeout="40s" />
-<action name="status" depth="0"  timeout="20s" interval="5s" />
-<action name="monitor" depth="0"  timeout="20s" interval="5s" />
+<action name="status"  timeout="20s" interval="60s" />
+<action name="monitor" depth="0" timeout="10s" interval="5s" />
+<action name="notify"  timeout="10s" />
 <action name="meta-data"  timeout="20s" />
 </actions>
 </resource-agent>
 END
 
-return $OCF_SUCCESS
+return ${OCF_SUCCESS}
 }
 
 ganesha_grace_usage() {
@@ -73,10 +82,10 @@ ganesha_grace_usage() {
 # Make sure meta-data and usage always succeed
 case $__OCF_ACTION in
 	meta-data)	ganesha_meta_data
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	usage|help)	ganesha_usage
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	*)
 			;;
@@ -84,81 +93,89 @@ esac
 
 ganesha_grace_start()
 {
-	local result=""
-	local resourcename=""
-	local deadserver=""
-	local tmpIFS=${IFS}
-	local pid_file="/var/run/ganesha.nfsd.pid"
-
-	# RHEL6 /etc/init.d/nfs-ganesha adds "-p /var/run/ganesha.nfsd.pid"
-	# RHEL7 systemd does not. Would be nicer if all distros used the
-	# same pid file.
-	if [ -e /usr/lib/systemd/system/nfs-ganesha.service ]; then
-		pid_file="/var/run/ganesha.pid"
+        rc=${OCF_ERR_GENERIC}
+	ocf_log debug "ganesha_grace_start()"
+	attr=$(attrd_updater -Q -n ${OCF_RESKEY_grace_active})
+
+	# Three possibilities:
+	# 1. There is no attribute at all and attr_updater returns
+	#    a zero length string. This happens when
+	#    ganesha_mon::monitor hasn't run at least once to set
+	#    the attribute. The assumption here is that the system
+	#    is coming up. We pretend, for now, that the node is
+	#    healthy, to allow the system to continue coming up.
+	#    It will cure itself in a few seconds
+	# 2. There is an attribute, and it has the value "1"; this
+	#    node is healthy.
+	# 3. There is an attribute, but it has no value or the value
+	#    "0"; this node is not healthy.
+
+	# case 1
+	if [[ -z "${attr}" ]]; then
+		return ${OCF_SUCCESS}
 	fi
 
-	# logger "ganesha_grace_start()"
-	# we're here because somewhere in the cluster one or more
-	# of the ganesha.nfsds have died, triggering a floating IP
-	# address to move. Resource constraint location rules ensure
-	# that this is invoked before the floating IP is moved.
-	if [ -e ${pid_file} -a \
-	     -d /proc/$(cat ${pid_file} ) ]; then
-		# my ganesha.nfsd is still running
-		# find out which one died?
-
-		pcs status | grep dead_ip-1 | sort > /tmp/.pcs_status
-
-		result=$(diff /var/run/ganesha/pcs_status /tmp/.pcs_status | grep '^>')
-		if [[ ${result} ]]; then
-			# logger "ganesha_grace_start(), ${result}"
-			IFS=$'\n'
-			for line in ${result}; do
-				resourcename=$(echo ${line} | cut -f 1 | cut -d ' ' -f 3)
-				deadserver=${resourcename%"-dead_ip-1"}
-
-				if [[ ${deadserver} ]]; then
-					# logger "ganesha_grace_start(), ${line}"
-					# logger "ganesha_grace_start(), dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/admin org.ganesha.nfsd.admin.grace string:${deadserver}"
-					dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/admin org.ganesha.nfsd.admin.grace string:${deadserver}
-					if [ $? -ne 0 ]; then
-						logger "warning: dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/admin org.ganesha.nfsd.admin.grace string:${deadserver} failed"
-					fi
-				fi
-			done
-			IFS=${tmpIFS}
-		fi
-
+	# case 2
+	if [[ "${attr}" = *"value=\"1\"" ]]; then
+		return ${OCF_SUCCESS}
 	fi
-	return $OCF_SUCCESS
+
+	# case 3
+	return ${OCF_NOT_RUNNING}
 }
 
 ganesha_grace_stop()
 {
 
-	# logger "ganesha_grace_stop()"
-	return $OCF_SUCCESS
+	ocf_log debug "ganesha_grace_stop()"
+	return ${OCF_SUCCESS}
+}
+
+ganesha_grace_notify()
+{
+	mode="${OCF_RESKEY_CRM_meta_notify_type}-${OCF_RESKEY_CRM_meta_notify_operation}"
+	case "$mode" in
+	post-stop)
+		ocf_log debug "stop_uname:${OCF_RESKEY_CRM_meta_notify_stop_uname}"
+		dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/admin org.ganesha.nfsd.admin.grace string:${OCF_RESKEY_CRM_meta_notify_stop_uname}
+		if [ $? -ne 0 ]; then
+			ocf_log info "dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/admin org.ganesha.nfsd.admin.grace string:${OCF_RESKEY_CRM_meta_notify_stop_uname} failed"
+		fi
+		;;
+	esac
+
+	return ${OCF_SUCCESS}
 }
 
 ganesha_grace_monitor()
 {
-	# logger "ganesha_grace_monitor()"
-	if [ ! -d /var/run/ganesha ]; then
-		mkdir -p /var/run/ganesha
+        rc=${OCF_ERR_GENERIC}
+	ocf_log debug "monitor"
+
+	attr=$(attrd_updater -Q -n ${OCF_RESKEY_grace_active})
+
+	# if there is no attribute (yet), maybe it's because
+	# this RA started before ganesha_mon (nfs-mon) has had
+	# chance to create it. In which case we'll pretend
+	# everything is okay this time around
+	if [[ -z "${attr}" ]]; then
+		return ${OCF_SUCCESS}
+	fi
+
+	if [[ "${attr}" = *"value=\"1\"" ]]; then
+		rc=${OCF_SUCCESS}
 	fi
-	pcs status | grep dead_ip-1 | sort > /var/run/ganesha/pcs_status
-	return $OCF_SUCCESS
+
+	return ${rc}
 }
 
 ganesha_grace_validate()
 {
-	return $OCF_SUCCESS
+	return ${OCF_SUCCESS}
 }
 
 ganesha_grace_validate
 
-# logger "ganesha_grace ${OCF_RESOURCE_INSTANCE} $__OCF_ACTION"
-
 # Translate each action into the appropriate function call
 case $__OCF_ACTION in
 start)          ganesha_grace_start
@@ -167,14 +184,16 @@ stop)           ganesha_grace_stop
 		;;
 status|monitor) ganesha_grace_monitor
 		;;
+notify)         ganesha_grace_notify
+		;;
 *)              ganesha_grace_usage
-                exit $OCF_ERR_UNIMPLEMENTED
+                exit ${OCF_ERR_UNIMPLEMENTED}
                 ;;
 esac
 
 rc=$?
 
 # The resource agent may optionally log a debug message
-ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION returned $rc"
+ocf_log debug "${OCF_RESOURCE_INSTANCE} ${__OCF_ACTION} returned $rc"
 exit $rc
 
diff --git a/extras/ganesha/ocf/ganesha_mon b/extras/ganesha/ocf/ganesha_mon
index c8e7de9c45e..f55cf7f2af3 100644
--- a/extras/ganesha/ocf/ganesha_mon
+++ b/extras/ganesha/ocf/ganesha_mon
@@ -29,14 +29,21 @@
 : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
 . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
 
-if [ -n "$OCF_DEBUG_LIBRARY" ]; then
-    . $OCF_DEBUG_LIBRARY
+if [ -n "${OCF_DEBUG_LIBRARY}" ]; then
+    . ${OCF_DEBUG_LIBRARY}
 else
     : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
 . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
 fi
 
-GRACE_DELAY=7
+# Defaults
+OCF_RESKEY_ganesha_active_default="ganesha-active"
+OCF_RESKEY_grace_active_default="grace-active"
+OCF_RESKEY_grace_delay_default="5"
+
+: ${OCF_RESKEY_ganesha_active=${OCF_RESKEY_ganesha_active_default}}
+: ${OCF_RESKEY_grace_active=${OCF_RESKEY_grace_active_default}}
+: ${OCF_RESKEY_grace_delay=${OCF_RESKEY_grace_delay_default}}
 
 ganesha_meta_data() {
         cat <<END
@@ -53,19 +60,37 @@ resource agent for nfs-ganesha.
 <shortdesc lang="en">Manages the user-space nfs-ganesha NFS server</shortdesc>
 
 <parameters>
+<parameter name="ganesha_active">
+<longdesc lang="en">NFS-Ganesha daemon active attribute</longdesc>
+<shortdesc lang="en">NFS-Ganesha daemon active attribute</shortdesc>
+<content type="string" default="ganesha-active" />
+</parameter>
+<parameter name="grace_active">
+<longdesc lang="en">NFS-Ganesha grace active attribute</longdesc>
+<shortdesc lang="en">NFS-Ganesha grace active attribute</shortdesc>
+<content type="string" default="grace-active" />
+</parameter>
+<parameter name="grace_delay">
+<longdesc lang="en">
+NFS-Ganesha grace delay.
+When changing this, adjust the ganesha_grace RA's monitor interval to match.
+</longdesc>
+<shortdesc lang="en">NFS-Ganesha grace delay</shortdesc>
+<content type="string" default="5" />
+</parameter>
 </parameters>
 
 <actions>
 <action name="start"   timeout="40s" />
 <action name="stop"    timeout="40s" />
-<action name="status" depth="0"  timeout="20s" interval="10s" />
+<action name="status"  timeout="20s" interval="60s" />
 <action name="monitor" depth="0"  timeout="10s" interval="10s" />
 <action name="meta-data"  timeout="20s" />
 </actions>
 </resource-agent>
 END
 
-return $OCF_SUCCESS
+return ${OCF_SUCCESS}
 }
 
 ganesha_mon_usage() {
@@ -73,12 +98,12 @@ ganesha_mon_usage() {
 }
 
 # Make sure meta-data and usage always succeed
-case $__OCF_ACTION in
+case ${__OCF_ACTION} in
 	meta-data)	ganesha_meta_data
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	usage|help)	ganesha_usage
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	*)
 			;;
@@ -86,12 +111,15 @@ esac
 
 ganesha_mon_start()
 {
-	return $OCF_SUCCESS
+	ocf_log debug "ganesha_mon_start"
+	ganesha_mon_monitor
+        return $OCF_SUCCESS
 }
 
 ganesha_mon_stop()
 {
-	return $OCF_SUCCESS
+	ocf_log debug "ganesha_mon_stop"
+        return $OCF_SUCCESS
 }
 
 ganesha_mon_monitor()
@@ -108,50 +136,56 @@ ganesha_mon_monitor()
 
 	if [ -e ${pid_file} -a \
 	     -d /proc/$(cat ${pid_file} ) ]; then
-		( pcs resource delete ${short_host}-dead_ip-1 > /dev/null 2>&1 )
 
-		attrd_updater -n ganesha-active -v 1
+		attrd_updater -n ${OCF_RESKEY_ganesha_active} -v 1
 		if [ $? -ne 0 ]; then
-			logger "warning: attrd_updater -n ganesha-active -v 1 failed"
+			ocf_log info "warning: attrd_updater -n ${OCF_RESKEY_ganesha_active} -v 1 failed"
 		fi
 
-	else
-		( pcs resource create ${short_host}-dead_ip-1 ocf:heartbeat:Dummy > /dev/null 2>&1 )
+		attrd_updater -n ${OCF_RESKEY_grace_active} -v 1
 		if [ $? -ne 0 ]; then
-			logger "warning: pcs resource create ${short_host}-dead_ip-1 ocf:heartbeat:Dummy failed"
+			ocf_log info "warning: attrd_updater -n ${OCF_RESKEY_grace_active} -v 1 failed"
 		fi
 
-		# The ${this-node}-dead_ip-1 resource is used to indicate
-		# that this ganesha.nfsd has died.
-		# VIP fail-over is then triggered by clearing the
-		# ganesha-active node attribute on this node.
-		#
-		# Meanwhile the ganesha_grace monitor() runs every 5
-		# seconds. We need to allow time for it to run and put
-		# the remaining ganesha.nfsds into grace before initiating
-		# the VIP fail-over.
-		sleep ${GRACE_DELAY}
-
-		attrd_updater -D -n ganesha-active
-		if [ $? -ne 0 ]; then
-			logger "warning: attrd_updater -D -n ganesha-active failed"
-		fi
+		return ${OCF_SUCCESS}
 	fi
 
-	return $OCF_SUCCESS
+	# VIP fail-over is triggered by clearing the
+	# ganesha-active node attribute on this node.
+	#
+	# Meanwhile the ganesha_grace notify() runs when its
+	# nfs-grace resource is disabled on a node; which
+	# is triggered by clearing the ganesha-grace node
+	# attribute on this node.
+	#
+	# We need to allow time for it to run and put
+	# the remaining ganesha.nfsds into grace before
+	# initiating the VIP fail-over.
+
+	attrd_updater -D -n ${OCF_RESKEY_grace_active}
+	if [ $? -ne 0 ]; then
+		ocf_log info "warning: attrd_updater -D -n ${OCF_RESKEY_grace_active} failed"
+	fi
+
+	sleep ${OCF_RESKEY_grace_delay}
+
+	attrd_updater -D -n ${OCF_RESKEY_ganesha_active}
+	if [ $? -ne 0 ]; then
+		ocf_log info "warning: attrd_updater -D -n ${OCF_RESKEY_ganesha_active} failed"
+	fi
+
+	return ${OCF_SUCCESS}
 }
 
 ganesha_mon_validate()
 {
-	return $OCF_SUCCESS
+	return ${OCF_SUCCESS}
 }
 
 ganesha_mon_validate
 
-# logger "ganesha_mon ${OCF_RESOURCE_INSTANCE} $__OCF_ACTION"
-
 # Translate each action into the appropriate function call
-case $__OCF_ACTION in
+case ${__OCF_ACTION} in
 start)          ganesha_mon_start
 		;;
 stop)           ganesha_mon_stop
@@ -159,13 +193,13 @@ stop)           ganesha_mon_stop
 status|monitor) ganesha_mon_monitor
 		;;
 *)              ganesha_mon_usage
-                exit $OCF_ERR_UNIMPLEMENTED
+                exit ${OCF_ERR_UNIMPLEMENTED}
                 ;;
 esac
 
 rc=$?
 
 # The resource agent may optionally log a debug message
-ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION returned $rc"
+ocf_log debug "${OCF_RESOURCE_INSTANCE} ${__OCF_ACTION} returned $rc"
 exit $rc
 
diff --git a/extras/ganesha/ocf/ganesha_nfsd b/extras/ganesha/ocf/ganesha_nfsd
index e064183daef..a9d3e4d860f 100644
--- a/extras/ganesha/ocf/ganesha_nfsd
+++ b/extras/ganesha/ocf/ganesha_nfsd
@@ -29,13 +29,16 @@
 : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
 . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
 
-if [ -n "$OCF_DEBUG_LIBRARY" ]; then
-    . $OCF_DEBUG_LIBRARY
+if [ -n "${OCF_DEBUG_LIBRARY}" ]; then
+    . ${OCF_DEBUG_LIBRARY}
 else
     : ${OCF_FUNCTIONS_DIR=${OCF_ROOT}/lib/heartbeat}
 . ${OCF_FUNCTIONS_DIR}/ocf-shellfuncs
 fi
 
+OCF_RESKEY_ha_vol_mnt_default="/var/run/gluster/shared_storage"
+: ${OCF_RESKEY_ha_vol_mnt=${OCF_RESKEY_ha_vol_mnt_default}}
+
 ganesha_meta_data() {
         cat <<END
 <?xml version="1.0"?>
@@ -59,16 +62,16 @@ resource agent for nfs-ganesha.
 </parameters>
 
 <actions>
-<action name="start"   timeout="40s" />
-<action name="stop"    timeout="40s" />
-<action name="status" depth="0"  timeout="20s" interval="1m" />
-<action name="monitor" depth="0"  timeout="10s" interval="1m" />
+<action name="start"   timeout="5s" />
+<action name="stop"    timeout="5s" />
+<action name="status" depth="0"  timeout="5s" interval="0" />
+<action name="monitor" depth="0"  timeout="5s" interval="0" />
 <action name="meta-data"  timeout="20s" />
 </actions>
 </resource-agent>
 END
 
-return $OCF_SUCCESS
+return ${OCF_SUCCESS}
 }
 
 ganesha_nfsd_usage() {
@@ -78,10 +81,10 @@ ganesha_nfsd_usage() {
 # Make sure meta-data and usage always succeed
 case $__OCF_ACTION in
 	meta-data)	ganesha_meta_data
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	usage|help)	ganesha_usage
-			exit $OCF_SUCCESS
+			exit ${OCF_SUCCESS}
 			;;
 	*)
 			;;
@@ -89,58 +92,60 @@ esac
 
 ganesha_nfsd_start()
 {
-	return $OCF_SUCCESS
+	local long_host=$(hostname)
+
+	if [[ -d /var/lib/nfs ]]; then
+		mv /var/lib/nfs /var/lib/nfs.backup
+		if [ $? -ne 0 ]; then
+			ocf_log notice "mv /var/lib/nfs /var/lib/nfs.backup failed"
+		fi
+		ln -s ${OCF_RESKEY_ha_vol_mnt}/nfs-ganesha/${long_host}/nfs /var/lib/nfs
+		if [ $? -ne 0 ]; then
+			ocf_log notice "ln -s ${OCF_RESKEY_ha_vol_mnt}/nfs-ganesha/${long_host}/nfs /var/lib/nfs failed"
+		fi
+	fi
+
+	return ${OCF_SUCCESS}
 }
 
 ganesha_nfsd_stop()
 {
-	local short_host=$(hostname -s)
-	local long_host=""
-
-	if [ "X${OCF_RESOURCE_INSTANCE:0:9}X" = "Xnfs_startX" ]; then
-
-		# if this is any nfs_start, go ahead. worst case we
-		# find the link already exists and do nothing
-		long_host=$(hostname)
-
-		if [ -d /var/lib/nfs ]; then
-			mv /var/lib/nfs /var/lib/nfs.backup
-			ln -s $OCF_RESKEY_ha_vol_mnt/nfs-ganesha/${long_host}/nfs /var/lib/nfs
-			if [ $? -ne 0 ]; then
-				logger "warning: ln -s $OCF_RESKEY_ha_vol_mnt/nfs-ganesha/${long_host}/nfs /var/lib/nfs failed"
-			fi
 
+	if [ -L /var/lib/nfs -a -d /var/lib/nfs.backup ]; then
+		rm -f /var/lib/nfs
+		if [ $? -ne 0 ]; then
+			ocf_log notice "rm -f /var/lib/nfs failed"
 		fi
-	else
-
-		# if this is a clone resource or is specific to this node
-		# remove the symlink and restore /var/lib/nfs
-
-		if [ "X${OCF_RESOURCE_INSTANCE}X" = "Xnfs_stopX" ] ||
-		   [ "X${OCF_RESOURCE_INSTANCE}X" = "Xnfs_stop-${short_host}X" ]; then
-			if [ -L /var/lib/nfs -a -d /var/lib/nfs.backup ]; then
-				rm -f /var/lib/nfs
-				mv /var/lib/nfs.backup /var/lib/nfs
-			fi
+		mv /var/lib/nfs.backup /var/lib/nfs
+		if [ $? -ne 0 ]; then
+			ocf_log notice "mv /var/lib/nfs.backup /var/lib/nfs failed"
 		fi
 	fi
 
-	return $OCF_SUCCESS
+	return ${OCF_SUCCESS}
 }
 
 ganesha_nfsd_monitor()
 {
-	return $OCF_SUCCESS
+	# pacemaker checks to see if RA is already running before starting it.
+	# if we return success, then it's presumed it's already running and
+	# doesn't need to be started, i.e. invoke the start action.
+	# return something other than success to make pacemaker invoke the
+	# start action
+	if [[ -L /var/lib/nfs ]]; then
+		return ${OCF_SUCCESS}
+	fi
+	return ${OCF_NOT_RUNNING}
 }
 
 ganesha_nfsd_validate()
 {
-	return $OCF_SUCCESS
+	return ${OCF_SUCCESS}
 }
 
 ganesha_nfsd_validate
 
-# logger "ganesha_nfsd ${OCF_RESOURCE_INSTANCE} $__OCF_ACTION"
+# ocf_log notice "ganesha_nfsd ${OCF_RESOURCE_INSTANCE} $__OCF_ACTION"
 
 # Translate each action into the appropriate function call
 case $__OCF_ACTION in
@@ -151,13 +156,13 @@ stop)           ganesha_nfsd_stop
 status|monitor) ganesha_nfsd_monitor
 		;;
 *)              ganesha_nfsd_usage
-                exit $OCF_ERR_UNIMPLEMENTED
+                exit ${OCF_ERR_UNIMPLEMENTED}
                 ;;
 esac
 
 rc=$?
 
 # The resource agent may optionally log a debug message
-ocf_log debug "${OCF_RESOURCE_INSTANCE} $__OCF_ACTION returned $rc"
+ocf_log debug "${OCF_RESOURCE_INSTANCE} ${__OCF_ACTION} returned $rc"
 exit $rc