From 42103b568698d8087d27f0848b402ccb7cfac86b Mon Sep 17 00:00:00 2001
From: Martin Schwenke <martin@meltin.net>
Date: Wed, 20 Mar 2019 17:35:44 +1100
Subject: [PATCH] ctdb-scripts: Stop/start mount/rquotad/status via NFS
 call-out

When an NFS check restarts a failed service by hand then systemd will
be unable to stop or start this service again because (at least) the
PID file will be wrong.  Do this via the NFS Linux kernel call-out
instead.  Allow the call-out to use the services instead of doing
manual restarts.  Add variables for mount, status and rquotad services
to support this.

Adding systemd NFS services to the call-out will follow.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13860

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@samba.org>
---
 ctdb/config/nfs-checks.d/10.status.check  |  4 +-
 ctdb/config/nfs-checks.d/40.mountd.check  |  4 +-
 ctdb/config/nfs-checks.d/50.rquotad.check |  4 +-
 ctdb/config/nfs-linux-kernel-callout      | 82 +++++++++++++++++++++++
 4 files changed, 88 insertions(+), 6 deletions(-)

diff --git a/ctdb/config/nfs-checks.d/10.status.check b/ctdb/config/nfs-checks.d/10.status.check
index dfa5c59117e..b8ce1e0ad0b 100644
--- a/ctdb/config/nfs-checks.d/10.status.check
+++ b/ctdb/config/nfs-checks.d/10.status.check
@@ -2,6 +2,6 @@
 version="1"
 restart_every=2
 unhealthy_after=6
-service_stop_cmd="killall -q -9 rpc.statd"
-service_start_cmd="rpc.statd ${STATD_HA_CALLOUT:+-H} $STATD_HA_CALLOUT ${STATD_HOSTNAME:+-n} $STATD_HOSTNAME ${STATD_PORT:+-p} $STATD_PORT ${STATD_OUTGOING_PORT:+-o} $STATD_OUTGOING_PORT"
+service_stop_cmd="$CTDB_NFS_CALLOUT stop status"
+service_start_cmd="$CTDB_NFS_CALLOUT start status"
 service_debug_cmd="program_stack_traces rpc.statd 5"
diff --git a/ctdb/config/nfs-checks.d/40.mountd.check b/ctdb/config/nfs-checks.d/40.mountd.check
index 56b3fd29512..bfe4c277ce9 100644
--- a/ctdb/config/nfs-checks.d/40.mountd.check
+++ b/ctdb/config/nfs-checks.d/40.mountd.check
@@ -2,6 +2,6 @@
 version="1"
 restart_every=2
 unhealthy_after=6
-service_stop_cmd="killall -q -9 rpc.mountd"
-service_start_cmd="rpc.mountd $RPCMOUNTDOPTS ${MOUNTD_PORT:+-p} $MOUNTD_PORT"
+service_stop_cmd="$CTDB_NFS_CALLOUT stop mountd"
+service_start_cmd="$CTDB_NFS_CALLOUT start mountd"
 service_debug_cmd="program_stack_traces rpc.mountd 5"
diff --git a/ctdb/config/nfs-checks.d/50.rquotad.check b/ctdb/config/nfs-checks.d/50.rquotad.check
index b7bd9d2c757..98bd8d98fce 100644
--- a/ctdb/config/nfs-checks.d/50.rquotad.check
+++ b/ctdb/config/nfs-checks.d/50.rquotad.check
@@ -2,6 +2,6 @@
 version="1"
 restart_every=2
 unhealthy_after=6
-service_stop_cmd="killall -q -9 rpc.rquotad"
-service_start_cmd="rpc.rquotad ${RQUOTAD_PORT:+-p} $RQUOTAD_PORT"
+service_stop_cmd="$CTDB_NFS_CALLOUT stop rquotad"
+service_start_cmd="$CTDB_NFS_CALLOUT start rquotad"
 service_debug_cmd="program_stack_traces rpc.rquotad 5"
diff --git a/ctdb/config/nfs-linux-kernel-callout b/ctdb/config/nfs-linux-kernel-callout
index ca7e2b88e96..74f7f0a16b0 100755
--- a/ctdb/config/nfs-linux-kernel-callout
+++ b/ctdb/config/nfs-linux-kernel-callout
@@ -21,7 +21,11 @@ sysvinit-*)
 	# Defaults
 	nfs_service="nfs"
 	nfs_lock_service=""
+	nfs_mountd_service=""
+	nfs_status_service=""
+	nfs_rquotad_service=""
 	nfs_config="/etc/sysconfig/nfs"
+	nfs_rquotad_config="$nfs_config"
 
 	case "$nfs_distro_style" in
 	*-redhat)
@@ -33,6 +37,7 @@ sysvinit-*)
 	*-debian)
 		nfs_service="nfs-kernel-server"
 		nfs_config="/etc/default/nfs-kernel-server"
+		nfs_rquotad_config="/etc/default/quota"
 		;;
 	*)
 		echo "Internal error"
@@ -134,6 +139,33 @@ service_stop ()
 	nlockmgr)
 		basic_stop "nfslock" >/dev/null 2>&1 || true
 		;;
+	mountd)
+		if [ -n "$nfs_mountd_service" ] ; then
+			service "$nfs_mountd_service" stop
+			return
+		fi
+
+		# Default to stopping by hand
+		killall -q -9 rpc.mountd
+		;;
+	rquotad)
+		if [ -n "$nfs_rquotad_service" ] ; then
+			service "$nfs_rquotad_service" stop
+			return
+		fi
+
+		# Default to stopping by hand
+		killall -q -9 rpc.rquotad
+		;;
+	status)
+		if [ -n "$nfs_status_service" ] ; then
+			service "$nfs_status_service" stop
+			return
+		fi
+
+		# Default to stopping by hand
+		killall -q -9 rpc.statd
+		;;
 	*)
 		usage
 	esac
@@ -148,6 +180,56 @@ service_start ()
 	nlockmgr)
 		basic_start "nfslock"
 		;;
+	mountd)
+		if [ -n "$nfs_mountd_service" ] ; then
+			service "$nfs_mountd_service" start
+			return
+		fi
+
+		# Default to starting by hand
+		nfs_load_config
+		if [ -z "$RPCMOUNTDOPTS" ] ; then
+			RPCMOUNTDOPTS="${MOUNTD_PORT:+-p }$MOUNTD_PORT"
+		fi
+		# shellcheck disable=SC2086
+		rpc.mountd $RPCMOUNTDOPTS
+		;;
+	rquotad)
+		if [ -n "$nfs_rquotad_service" ] ; then
+			service "$nfs_rquotad_service" start
+			return
+		fi
+
+		# Default to starting by hand
+		nfs_load_config "$nfs_rquotad_config"
+		if [ -z "$RPCRQUOTADOPTS" ] ; then
+			RPCRQUOTADOPTS="${RQUOTAD_PORT:+-p }$RQUOTAD_PORT"
+		fi
+		# shellcheck disable=SC2086
+		rpc.rquotad $RPCRQUOTADOPTS
+		;;
+	status)
+		if [ -n "$nfs_status_service" ] ; then
+			service "$nfs_status_service" start
+			return
+		fi
+
+		# Default to starting by hand
+		nfs_load_config
+		# Red Hat uses STATDARG, Debian uses STATDOPTS
+		opts="${STATDARG:-${STATDOPTS:-''}}"
+		if [ -z "$opts" ] ; then
+			# shellcheck disable=SC2086
+			set -- \
+			    ${STATD_HA_CALLOUT:+-H} $STATD_HA_CALLOUT \
+			    ${STATD_HOSTNAME:+-n} $STATD_HOSTNAME \
+			    ${STATD_PORT:+-p} $STATD_PORT \
+			    ${STATD_OUTGOING_PORT:+-o} $STATD_OUTGOING_PORT
+			opts="$*"
+		fi
+		# shellcheck disable=SC2086
+		rpc.statd $opts
+		;;
 	*)
 		usage
 	esac