RHEL7 was adopted Pacemaker.
RHEL5 or RHEL 6 was used RGMANAGER.
| Formation | Note |
|---|---|
| Pacemaker + Corosync | recommend |
| Pacemaker + Heartbeat Versinon3 |
| Package | Note |
|---|---|
| pacemaker | Pacemaker is an Open Source, High Availability resource manager suitable for both small and large clusters. |
| corosync | cluster.conf |
| pcs | a full cluster lifecycle configuration shell and web based GUI. |
You must start pacemaker for pcs commands.
| Commands | Operation | Note |
|---|---|---|
| pcs status pcs status --full crm_mon -1 | Check Status | |
| pcs config | Check Configuration | |
| pcs property list | Check Cluster Properties | |
| pcs cluster cib | Check Configuration with XML | |
| pcs status nodes | Check Note Status |
| Commands | Operation | Note |
|---|---|---|
| pcs cluster start --all | start cluster with two node | start pacemaker and corosync |
| pcs cluster start SERVER1 | start cluster with one node, SERVER1 | start pacemaker and corosync |
| pcs cluster stop --all | stop cluster | stop pacemaker and corosync |
| pcs cluster stop SERVER1 | stop cluster with one node, SERVER1 | stop pacemaker and corosync |
| pcs cluster enalbe --all | auto start | |
| pcs cluster stop SERVER1 --force | stop cluster on only one node | |
| pcs cluster kill | Forcibly stop cluster | |
| pcs cluster standby SERVER1 | change to standby | Cluster SERVER1 to SERVER2 |
| pcs cluster unstandby SERVER1 | change to unstandby |
| Commands | Operation | Note |
|---|---|---|
| pcs disalbe XXXX | stop resource | |
| pcs enable XXXX | start resource |
| Commands | Operation | Note |
|---|---|---|
| pcs cluster sync | Sync corosync.conf | |
| pcs cluster destroy | Delete Configuration | Deleted /var/lib/pacemaker/cib/cib.xml, too. Deleted /etc/cluster/cluster.conf, too. |
| pcs cluster cib output.cib | Output Configuration | |
| pcs -f output.cib COMMAND | Change Configuration | |
| pcs cluster cib-push output.cib | Input Configuration |
| Commands | Operation |
|---|---|
| pcs resource group list | The following command lists all currently configured resource groups |
| pcs resource group add group_name resource_id | You create a resource group with the following command If the group does not exist, this command creates the group. If the group exists, this command adds additional resources to the group. |
| pcs resource group remove group_name resource_id… | You remove a resource from a group with the following command. If there are no resources in the group, this command removes the group itself. |
| pcs resource show | |
| pcs resource group list | |
| pcs resource show RESCOURCE | |
| pcs cluster corosync | check node and totem |
| pcs stonith show RESOURCE | |
| pcs constraint list | check the resource order |
node1/2# yum install pacemaker corsync node1/2# yum install pcs fence-agents-all
node1/2# systemctl enable pcsd node1/2# systemctl start pcsd
node1/2# echo redhat |passwd --stdin hacluster
node1# pcs cluster auth -u hacluster -p redhat server1 server2
node1# pcs cluster setup --start --name test-cluster --wait_for_all=0 server1 server2 --wait_for_all=0 : If it is 1node, start cluster.
node1# pcs status
# pcs property list Cluster Properties: cluster-infrastructure: corosync cluster-name: test-cluster dc-version: 1.1.16-a14efad default-resource-stickiness: INFINITY have-watchdog: false no-quorum-policy: ignore stonith-enabled: false
# pcs property set stonith-enabled=false
# pcs property set default-resource-stickiness="INFINITY"
Even if the number of clusters has not reached the required minimum number, no special action is taken and management of all nodes is continued.
# pcs property set no-quorum-policy=ignore
STONITH (Shoot The Other Node In The Head)
#pcs stonith list #pcs stonith describe XXXXX
#pcs stonith list #pcs stonith describe XXXXX
## unique name fence agent target node device addr options # pcs stonith create fence_server1 fence_ilo4 pcmk_host_list="server2-ilo" ipaddr="xx.xx.xx.xx" action="reboot" login="admin" passwd="secret" delay=15 op monitor interval=60s # pcs stonith create fence_server2 fence_ilo4 pcmk_host_list="server2-ilo" ipaddr="xx.xx.xx.xx" action="reboot" login="admin" passwd="secret" op monitor interval=60s If you dont't use Fence, You can do action="off".
# pcs stonith show stonith-ilo4-server1 (stonith:fence_ilo4): Started stonith-ilo4-server2 (stonith:fence_ilo4): Started
#pcs property --all |grep -E "stonith-enabled|stonith-action" stonith-action: reboot stonith-enabled: true
# pcs stonith fence server2 -> server2 is rebooted.
# pcs resource disable RESOURCE_STONITH # pcs resource enable RESOURCE_STONITH
disable auto failover
It is failovered with manual operation or node down.
'migration-threshold' is retry number of resource restart.
pcs resource defaults resource-stickiness="INFINITY" migration-threshold="0"
set virtual IP address
pcs resource create ipfront ocf:heartbeat:IPaddr2 ip=192.168.40.xx cidr_netmask=24 nic="eth0" op monitor interval=10s pcs resource create ipmaintenance ocf:heartbeat:IPaddr2 ip=192.168.50.xx cidr_netmask=24 nic="eth1" op monitor interval=10s
pcs resource create httpd systemd:httpd --group zabbix-group pcs resource create php-fpm systemd:php-fpm --group zabbix-group pcs resource create zabbix-server systemd:zabbix-server --group zabbix-group
/etc/init.d/mysqld
pcs resource create mysql lsb:mysqld op monitor interval=60s pcs resource update mysql op start interval=30s timeout=15 on-fail=ignore pcs resource update mysql op stop interval=0s timeout=15 on-fail=ignore
pcs constraint order frontip then maintenanceip pcs constraint order maintenanceip then mysql
pcs resource group add mysqlgroup frontip maintenanceip mysql
# pcs resouce delete XXXX
# pcs resource cleanup res_stonith_ipmilan01
# pcs config backup /tmp/`date +%Y%m%d`_`uname -n`_pcs_config_backup # pcs config restore /tmp/`date +%Y%m%d`_`uname -n`_pcs_config_backup.tar.bz2
/etc/corosync/corosync.conf
/etc/corosync/corosync.conf.example
In AWS EC2, You must use UDP Unicast insted of UDP Multicast in corosync.
Sample Configuration : /etc/corosync/corosync.conf.example.udpu