Adding upstream version 4.2.

Signed-off-by: Daniel Baumann <daniel@debian.org>

clustermd_tests/00r10_Create

@@ -0,0 +1,50 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check $NODE1 resync
+check $NODE2 PENDING
+check all wait
+check all raid10
+check all bitmap
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l10 -b clustered -n3 --layout n3 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid10
+check all bitmap
+check all state UUU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l10 -b clustered -n2 -x1 --layout n2 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid10
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+name=tstmd
+mdadm -CR $md0 -l10 -b clustered -n2 $dev0 $dev1 --layout n2 --name=$name --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+for ip in $NODE1 $NODE2
+do
+	ssh $ip "mdadm -D $md0 | grep 'Name' | grep -q $name"
+	[ $? -ne '0' ] &&
+		die "$ip: check --name=$name failed."
+done
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/00r1_Create

@@ -0,0 +1,50 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check $NODE1 resync
+check $NODE2 PENDING
+check all wait
+check all raid1
+check all bitmap
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid1
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+name=tstmd
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --name=$name --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+for ip in $NODE1 $NODE2
+do
+	ssh $ip "mdadm -D $md0 | grep 'Name' | grep -q $name"
+	[ $? -ne '0' ] &&
+		die "$ip: check --name=$name failed."
+done
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/01r10_Grow_bitmap-switch

@@ -0,0 +1,51 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+
+# switch 'clustered' bitmap to 'none', and then 'none' to 'internal'
+stop_md $NODE2 $md0
+mdadm --grow $md0 --bitmap=none
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'clustered' to 'none' failed."
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] &&
+	die "$NODE1: bitmap still exists in member_disks."
+check all nobitmap
+mdadm --grow $md0 --bitmap=internal
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'none' to 'internal' failed."
+sleep 1
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] ||
+	die "$NODE1: create 'internal' bitmap failed."
+check $NODE1 bitmap
+
+# switch 'internal' bitmap to 'none', and then 'none' to 'clustered'
+mdadm --grow $md0 --bitmap=none
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'internal' to 'none' failed."
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] &&
+	die "$NODE1: bitmap still exists in member_disks."
+check $NODE1 nobitmap
+mdadm --grow $md0 --bitmap=clustered
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'none' to 'clustered' failed."
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+sleep 1
+for ip in $NODES
+do
+	ssh $ip "mdadm -X $dev0 $dev1 | grep -q 'Cluster name'" ||
+		die "$ip: create 'clustered' bitmap failed."
+done
+check all bitmap
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/01r10_Grow_resize

@@ -0,0 +1,38 @@
+#!/bin/bash
+
+size=20000
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 --size $size --chunk=64 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+
+mdadm --grow $md0 --size max
+check $NODE1 resync
+check $NODE1 wait
+check all state UU
+
+mdadm --grow $md0 --size $size
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 --chunk=64 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+
+mdadm --grow $md0 --chunk=128
+check $NODE1 reshape
+check $NODE1 wait
+check all chunk 128
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/01r1_Grow_add

@@ -0,0 +1,68 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --grow $md0 --raid-devices=3 --add $dev2
+sleep 0.3
+grep recovery /proc/mdstat
+if [ $? -eq '0' ]
+then
+	check $NODE1 wait
+else
+	check $NODE2 recovery
+	check $NODE2 wait
+fi
+check all state UUU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid1
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --grow $md0 --raid-devices=3 --add $dev3
+sleep 0.3
+grep recovery /proc/mdstat
+if [ $? -eq '0' ]
+then
+	check $NODE1 wait
+else
+	check $NODE2 recovery
+	check $NODE2 wait
+fi
+check all state UUU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid1
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --grow $md0 --raid-devices=3
+sleep 0.3
+grep recovery /proc/mdstat
+if [ $? -eq '0' ]
+then
+	check $NODE1 wait
+else
+	check $NODE2 recovery
+	check $NODE2 wait
+fi
+check all state UUU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/01r1_Grow_bitmap-switch

@@ -0,0 +1,51 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+
+# switch 'clustered' bitmap to 'none', and then 'none' to 'internal'
+stop_md $NODE2 $md0
+mdadm --grow $md0 --bitmap=none
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'clustered' to 'none' failed."
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] &&
+	die "$NODE1: bitmap still exists in member_disks."
+check all nobitmap
+mdadm --grow $md0 --bitmap=internal
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'none' to 'internal' failed."
+sleep 2
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] ||
+	die "$NODE1: create 'internal' bitmap failed."
+check $NODE1 bitmap
+
+# switch 'internal' bitmap to 'none', and then 'none' to 'clustered'
+mdadm --grow $md0 --bitmap=none
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'internal' to 'none' failed."
+mdadm -X $dev0 $dev1 &> /dev/null
+[ $? -eq '0' ] &&
+	die "$NODE1: bitmap still exists in member_disks."
+check $NODE1 nobitmap
+mdadm --grow $md0 --bitmap=clustered
+[ $? -eq '0' ] ||
+	die "$NODE1: change bitmap 'none' to 'clustered' failed."
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+sleep 2
+for ip in $NODES
+do
+	ssh $ip "mdadm -X $dev0 $dev1 | grep -q 'Cluster name'" ||
+		die "$ip: create 'clustered' bitmap failed."
+done
+check all bitmap
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/01r1_Grow_resize

@@ -0,0 +1,23 @@
+#!/bin/bash
+
+size=10000
+
+mdadm -CR $md0 -l1 -b clustered --size $size -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+
+mdadm --grow $md0 --size max
+check $NODE1 resync
+check $NODE1 wait
+check all state UU
+
+mdadm --grow $md0 --size $size
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r10_Manage_add

@@ -0,0 +1,33 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0 --remove $dev0
+mdadm --zero $dev2
+mdadm --manage $md0 --add $dev2
+sleep 0.3
+check $NODE1 recovery
+check $NODE1 wait
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add $dev2
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r10_Manage_add-spare

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add-spare $dev2
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid10
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add-spare $dev3
+check all spares 2
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r10_Manage_re-add

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid10
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0 --remove $dev0
+mdadm --manage $md0 --re-add $dev0
+check $NODE1 recovery
+check all wait
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r1_Manage_add

@@ -0,0 +1,33 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0 --remove $dev0
+mdadm --zero $dev2
+mdadm --manage $md0 --add $dev2
+sleep 0.3
+check $NODE1 recovery
+check $NODE1 wait
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add $dev2
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r1_Manage_add-spare

@@ -0,0 +1,30 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add-spare $dev2
+check all spares 1
+check all state UU
+check all dmesg
+stop_md all $md0
+
+mdadm -CR $md0 -l1 -b clustered -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid1
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --manage $md0 --add-spare $dev3
+check all spares 2
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/02r1_Manage_re-add

@@ -0,0 +1,16 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check all nosync
+check all raid1
+check all bitmap
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0 --remove $dev0
+mdadm --manage $md0 --re-add $dev0
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/03r10_switch-recovery

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid10
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0
+sleep 0.2
+check $NODE1 recovery
+stop_md $NODE1 $md0
+check $NODE2 recovery
+check $NODE2 wait
+check $NODE2 state UU
+check all dmesg
+stop_md $NODE2 $md0
+
+exit 0

clustermd_tests/03r10_switch-resync

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l10 -b clustered --layout n2 -n2 $dev0 $dev1
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check $NODE1 resync
+check $NODE2 PENDING
+stop_md $NODE1 $md0
+check $NODE2 resync
+check $NODE2 wait
+mdadm -A $md0 $dev0 $dev1
+check all raid10
+check all bitmap
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/03r1_switch-recovery

@@ -0,0 +1,21 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 -x1 $dev0 $dev1 $dev2 --assume-clean
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1 $dev2
+check all nosync
+check all raid1
+check all bitmap
+check all spares 1
+check all state UU
+check all dmesg
+mdadm --manage $md0 --fail $dev0
+sleep 0.3
+check $NODE1 recovery
+stop_md $NODE1 $md0
+check $NODE2 recovery
+check $NODE2 wait
+check $NODE2 state UU
+check all dmesg
+stop_md $NODE2 $md0
+
+exit 0

clustermd_tests/03r1_switch-resync

@@ -0,0 +1,18 @@
+#!/bin/bash
+
+mdadm -CR $md0 -l1 -b clustered -n2 $dev0 $dev1
+ssh $NODE2 mdadm -A $md0 $dev0 $dev1
+check $NODE1 resync
+check $NODE2 PENDING
+stop_md $NODE1 $md0
+check $NODE2 resync
+check $NODE2 wait
+mdadm -A $md0 $dev0 $dev1
+check all raid1
+check all bitmap
+check all nosync
+check all state UU
+check all dmesg
+stop_md all $md0
+
+exit 0

clustermd_tests/cluster_conf

@@ -0,0 +1,43 @@
+# Prerequisite:
+# 1. The clustermd_tests/ cases only support to test 2-node-cluster, cluster
+#    requires packages: 'pacemaker+corosync+sbd+crmsh', all packages link at
+#    "https://github.com/ClusterLabs/", and also requires dlm resource running
+#    on each node of cluster.
+#    For quick start HA-cluster with SUSE distributions, refer to the chapter 6-8:
+#    https://www.suse.com/documentation/sle-ha-12/install-quick/data/install-quick.html
+#    For Redhat distributions, please refer to:
+#    https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html-single/high_availability_add-on_administration/index
+# 2. Setup ssh-access with no-authorized mode, it should be:
+#    # 'ssh $node1 -l root ls' and 'ssh $node2 -l root ls' success on any node.
+# 3. Fill-up node-ip part and disks part as following.
+
+# Set node1 as the master node, the cluster-md cases should run on this node,
+# and node2 is the slave node.
+# For example:
+# NODE1=192.168.1.100 (testing run here)
+# NODE2=192.168.1.101
+NODE1=
+NODE2=
+
+# Provide the devlist for clustermd-testing, alternative: if set the step 1,
+# don't set step 2, and vice versa.
+# 1. Use ISCSI service to provide shared storage, then login ISCSI target via
+#    to ISCSI_TARGET_ID and ISCSI_TARGET_IP on iscsi clients, commands like:
+#    Execute on iscsi clients:
+#    1) discover the iscsi server.
+#    # iscsiadm -m discovery -t st -p $ISCSI_TARGET_IP
+#    2) login and establish connection.
+#    # iscsiadm -m node -T $ISCSI_TARGET_ID -p $ISCSI_TARGET_IP -l
+# Note:
+#    On ISCSI server, must create all iscsi-luns in one target_id, recommend more
+#    than 6 luns/disks for testing, and each disk should be: 100M < disk < 800M.
+# 2. If all cluster-nodes mounted the same disks directly, and the devname are
+#    the same on all nodes, then put them to 'devlist'.
+
+# For example: (Only set $ISCSI_TARGET_ID is enough if iscsi has already connected)
+# ISCSI_TARGET_ID=iqn.2018-01.example.com:clustermd-testing
+# ISCSI_TARGET_IP=192.168.1.102
+ISCSI_TARGET_ID=
+
+#devlist=/dev/sda /dev/sdb /dev/sdc /dev/sdd
+devlist=

clustermd_tests/func.sh

@@ -0,0 +1,332 @@
+#!/bin/bash
+
+check_ssh()
+{
+	NODE1="$(grep '^NODE1' $CLUSTER_CONF | cut -d'=' -f2)"
+	NODE2="$(grep '^NODE2' $CLUSTER_CONF | cut -d'=' -f2)"
+	[ -z "$NODE1" -o -z "$NODE2" ] && {
+		echo "Please provide node-ip in $CLUSTER_CONF."
+		exit 1
+	}
+	for ip in $NODE1 $NODE2
+	do
+		ssh -o NumberOfPasswordPrompts=0 $ip -l root "pwd" > /dev/null
+		[ $? -ne 0 ] && {
+			echo "Please setup ssh-access with no-authorized mode."
+			exit 1
+		}
+	done
+}
+
+fetch_devlist()
+{
+	ISCSI_ID="$(grep '^ISCSI_TARGET_ID' $CLUSTER_CONF | cut -d'=' -f2)"
+	devlist="$(grep '^devlist' $CLUSTER_CONF | cut -d'=' -f2)"
+	if [ ! -z "$ISCSI_ID" -a ! -z "$devlist" ]
+	then
+		echo "Config ISCSI_TARGET_ID or devlist in $CLUSTER_CONF."
+		exit 1
+	elif [ ! -z "$ISCSI_ID" -a -z "$devlist" ]
+	then
+		for ip in $NODE1 $NODE2
+		do
+			ssh $ip "ls /dev/disk/by-path/*$ISCSI_ID*" > /dev/null
+			[ $? -ne 0 ] && {
+				echo "$ip: No disks found in '$ISCSI_ID' connection."
+				exit 1
+			}
+		done
+		devlist=($(ls /dev/disk/by-path/*$ISCSI_ID*))
+	fi
+	# sbd disk cannot use in testing
+	# Init devlist as an array
+	i=''
+	devlist=(${devlist[@]#$i})
+	for i in ${devlist[@]}
+	do
+		sbd -d $i dump &> /dev/null
+		[ $? -eq '0' ] && devlist=(${devlist[@]#$i})
+	done
+	for i in $(seq 0 ${#devlist[@]})
+	do
+		eval "dev$i=${devlist[$i]}"
+	done
+	[ "${#devlist[@]}" -lt 6 ] && {
+		echo "Cluster-md testing requires 6 disks at least."
+		exit 1
+	}
+}
+
+check_dlm()
+{
+	if ! crm configure show | grep -q dlm
+	then
+		crm configure primitive dlm ocf:pacemaker:controld \
+			op monitor interval=60 timeout=60 \
+			meta target-role=Started &> /dev/null
+		crm configure group base-group dlm
+		crm configure clone base-clone base-group \
+			meta interleave=true
+	fi
+	sleep 1
+	for ip in $NODE1 $NODE2
+	do
+		ssh $ip "pgrep dlm_controld > /dev/null" || {
+			echo "$ip: dlm_controld daemon doesn't exist."
+			exit 1
+		}
+	done
+	crm_mon -r -n1 | grep -iq "fail\|not" && {
+		echo "Please clear cluster-resource errors."
+		exit 1
+	}
+}
+
+check_env()
+{
+	user=$(id -un)
+	[ "X$user" = "Xroot" ] || {
+		echo "testing can only be done as 'root'."
+		exit 1
+	}
+	[ \! -x $mdadm ] && {
+		echo "test: please run make everything before perform testing."
+		exit 1
+	}
+	check_ssh
+	commands=(mdadm iscsiadm bc modinfo dlm_controld
+		  udevadm crm crm_mon lsblk pgrep sbd)
+	for ip in $NODE1 $NODE2
+	do
+		for cmd in ${commands[@]}
+		do
+			ssh $ip "which $cmd &> /dev/null" || {
+				echo "$ip: $cmd, command not found!"
+				exit 1
+			}
+		done
+		mods=(raid1 raid10 md_mod dlm md-cluster)
+		for mod in ${mods[@]}
+		do
+			ssh $ip "modinfo $mod > /dev/null" || {
+				echo "$ip: $mod, module doesn't exist."
+				exit 1
+			}
+		done
+		ssh $ip "lsblk -a | grep -iq raid"
+		[ $? -eq 0 ] && {
+			echo "$ip: Please run testing without running RAIDs environment."
+			exit 1
+		}
+		ssh $ip "modprobe md_mod"
+	done
+	fetch_devlist
+	check_dlm
+	[ -d $logdir ] || mkdir -p $logdir
+}
+
+# $1/node, $2/optional
+stop_md()
+{
+	if [ "$1" == "all" ]
+	then
+		NODES=($NODE1 $NODE2)
+	elif [ "$1" == "$NODE1" -o "$1" == "$NODE2" ]
+	then
+		NODES=$1
+	else
+		die "$1: unknown parameter."
+	fi
+	if [ -z "$2" ]
+	then
+		for ip in ${NODES[@]}
+		do
+			ssh $ip mdadm -Ssq
+		done
+	else
+		for ip in ${NODES[@]}
+		do
+			ssh $ip mdadm -S $2
+		done
+	fi
+}
+
+# $1/optional, it shows why to save log
+save_log()
+{
+	status=$1
+	logfile="$status""$_basename".log
+
+	cat $targetdir/stderr >> $targetdir/log
+	cp $targetdir/log $logdir/$_basename.log
+
+	for ip in $NODE1 $NODE2
+	do
+		echo "##$ip: saving dmesg." >> $logdir/$logfile
+		ssh $ip "dmesg -c" >> $logdir/$logfile
+		echo "##$ip: saving proc mdstat." >> $logdir/$logfile
+		ssh $ip "cat /proc/mdstat" >> $logdir/$logfile
+		array=($(ssh $ip "mdadm -Ds | cut -d' ' -f2"))
+
+		if [ ! -z "$array" -a ${#array[@]} -ge 1 ]
+		then
+			echo "##$ip: mdadm -D ${array[@]}" >> $logdir/$logfile
+			ssh $ip "mdadm -D ${array[@]}" >> $logdir/$logfile
+			md_disks=($(ssh $ip "mdadm -DY ${array[@]} | grep "/dev/" | cut -d'=' -f2"))
+			cat /proc/mdstat | grep -q "bitmap"
+			if [ $? -eq 0 ]
+			then
+				echo "##$ip: mdadm -X ${md_disks[@]}" >> $logdir/$logfile
+				ssh $ip "mdadm -X ${md_disks[@]}" >> $logdir/$logfile
+				echo "##$ip: mdadm -E ${md_disks[@]}" >> $logdir/$logfile
+				ssh $ip "mdadm -E ${md_disks[@]}" >> $logdir/$logfile
+			fi
+		else
+			echo "##$ip: no array assembled!" >> $logdir/$logfile
+		fi
+	done
+	[ "$1" == "fail" ] &&
+		echo "See $logdir/$_basename.log and $logdir/$logfile for details"
+	stop_md all
+}
+
+do_setup()
+{
+	check_env
+	ulimit -c unlimited
+}
+
+do_clean()
+{
+	for ip in $NODE1 $NODE2
+	do
+		ssh $ip "mdadm -Ssq; dmesg -c > /dev/null"
+	done
+	mdadm --zero ${devlist[@]} &> /dev/null
+}
+
+cleanup()
+{
+	check_ssh
+	do_clean
+}
+
+# check: $1/cluster_node $2/feature $3/optional
+check()
+{
+	NODES=()
+	if [ "$1" == "all" ]
+	then
+		NODES=($NODE1 $NODE2)
+	elif [ "$1" == "$NODE1" -o "$1" == "$NODE2" ]
+	then
+		NODES=$1
+	else
+		die "$1: unknown parameter."
+	fi
+	case $2 in
+		spares )
+			for ip in ${NODES[@]}
+			do
+				spares=$(ssh $ip "tr '] ' '\012\012' < /proc/mdstat | grep -c '(S)'")
+				[ "$spares" -ne "$3" ] &&
+					die "$ip: expected $3 spares, but found $spares"
+			done
+		;;
+		raid* )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -sq "$2" /proc/mdstat" ||
+					die "$ip: check '$2' failed."
+			done
+		;;
+		PENDING | recovery | resync | reshape )
+			cnt=5
+			for ip in ${NODES[@]}
+			do
+				while ! ssh $ip "grep -sq '$2' /proc/mdstat"
+				do
+					if [ "$cnt" -gt '0' ]
+					then
+						sleep 0.2
+						cnt=$[cnt-1]
+					else
+						die "$ip: no '$2' happening!"
+					fi
+				done
+			done
+		;;
+		wait )
+			local cnt=60
+			for ip in ${NODES[@]}
+			do
+				p=$(ssh $ip "cat /proc/sys/dev/raid/speed_limit_max")
+				ssh $ip "echo 200000 > /proc/sys/dev/raid/speed_limit_max"
+				while ssh $ip "grep -Esq '(resync|recovery|reshape|check|repair)' /proc/mdstat"
+				do
+					if [ "$cnt" -gt '0' ]
+					then
+						sleep 5
+						cnt=$[cnt-1]
+					else
+						die "$ip: Check '$2' timeout over 300 seconds."
+					fi
+				done
+				ssh $ip "echo $p > /proc/sys/dev/raid/speed_limit_max"
+			done
+		;;
+		bitmap )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -sq '$2' /proc/mdstat" ||
+					die "$ip: no '$2' found in /proc/mdstat."
+			done
+		;;
+		nobitmap )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -sq 'bitmap' /proc/mdstat" &&
+					die "$ip: 'bitmap' found in /proc/mdstat."
+			done
+		;;
+		chunk )
+			for ip in ${NODES[@]}
+			do
+				chunk_size=`awk -F',' '/chunk/{print $2}' /proc/mdstat | awk -F'[a-z]' '{print $1}'`
+				[ "$chunk_size" -ne "$3" ] &&
+					die "$ip: chunksize should be $3, but it's $chunk_size"
+			done
+		;;
+		state )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -Esq 'blocks.*\[$3\]\$' /proc/mdstat" ||
+					die "$ip: no '$3' found in /proc/mdstat."
+			done
+		;;
+		nosync )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -Eq '(resync|recovery)' /proc/mdstat" &&
+					die "$ip: resync or recovery is happening!"
+			done
+		;;
+		readonly )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "grep -sq "read-only" /proc/mdstat" ||
+					die "$ip: check '$2' failed!"
+			done
+		;;
+		dmesg )
+			for ip in ${NODES[@]}
+			do
+				ssh $ip "dmesg | grep -iq 'error\|call trace\|segfault'" &&
+					die "$ip: check '$2' prints errors!"
+			done
+		;;
+		* )
+			die "unknown parameter $2"
+		;;
+	esac
+}