Restart perd, eperd, eooqd if they die.

bin/ATLAS

@@ -91,6 +91,7 @@ $KILL_PERDS_CMD
 
 # the system part is done. now set env for ATLAS.
 rm -f $STATUS_DIR/*.vol
+rm -f $RUN_DIR/*.vol
 
 # Crontabs may be bad enough that we can't actually do anything. Delete them
 # after too many failed boots
@@ -136,9 +137,23 @@ mkdir -p $DATA_DIR/oneoff
 $CHOWN_FOR_MSM
 $CHMOD_FOR_MSM
 
-perd -c $BASE_DIR/crons/main -A 9801 -P $RUN_DIR/perd-main.pid.vol
-eperd -c $BASE_DIR/crons/7 -A 9807 -P $RUN_DIR/perd-7.pid.vol -O /home/atlas/data/new/7 -i 7
-eooqd $BASE_DIR/crons/oneoff -A 9809 -P $RUN_DIR/eooqd.pid.vol -i 9 &
+start_perd()
+{
+	perd -c $BASE_DIR/crons/main -A 9801 -P $RUN_DIR/perd-main.pid.vol
+}
+start_eperd()
+{
+	eperd -c $BASE_DIR/crons/7 -A 9807 -P $RUN_DIR/perd-7.pid.vol -O /home/atlas/data/new/7 -i 7
+}
+start_eooqd()
+{
+	eooqd $BASE_DIR/crons/oneoff -A 9809 -P $RUN_DIR/eooqd.pid.vol -i 9 &
+}
+
+start_perd
+start_eperd
+start_eooqd
+
 
 $NTPCLIENT_CMD &
 
@@ -295,6 +310,34 @@ do
 		$RESOLVCONF_CMD
 	fi
 
+	# Check if perd is still running
+	if check_pid $(cat $RUN_DIR/perd-main.pid.vol)
+	then
+		: # perd is running
+	else
+		echo restarting perd
+		start_perd
+		echo "RESULT 9801 perd died (restarted)" >> $DATA_NEW_DIR/simpleping
+	fi
+	# Check if eperd is still running
+	if check_pid $(cat $RUN_DIR/perd-7.pid.vol)
+	then
+		: # eperd is running
+	else
+		echo restarting eperd
+		start_eperd
+		echo "RESULT 9807 eperd died (restarted)" >> $DATA_NEW_DIR/simpleping
+	fi
+	# Check if eooqd is still running
+	if check_pid $(cat $RUN_DIR/eooqd.pid.vol)
+	then
+		: # eooqd is running
+	else
+		echo restarting eooqd
+		start_eooqd
+		echo "RESULT 9809 eooqd died (restarted)" >> $DATA_NEW_DIR/simpleping
+	fi
+
 	sync
 	$DROP_CACHES
 done

bin/arch/linux/linux-functions.sh

@@ -13,6 +13,10 @@ rchoose()
 {
 	$BB_BIN_DIR/rchoose "$@"
 }
+check_pid()
+{
+	$SU_CMD kill -0 "$@"
+}
 condmv()
 {
 	$BB_BIN_DIR/condmv "$@"