Pacemaker with GFS2 RHEL 7
Install and initial cluster configuration
/etc/hosts configuration example on both nodes:
192.168.122.10 cluster-1.exmaple.com cluster-1
192.168.122.15 cluster-2.example.com cluster-2
Yum configuration on both nodes:
[cdrom] name=cdrom baseurl=file:///media gpgcheck=0 enabled=1 [ha] name=ha baseurl=file:///media/addons/HighAvailability/ gpgcheck=0 enabled=1 [storage] name=storage baseurl=file:///media/addons/ResilientStorage/ gpgcheck=0 enabled=1
On both nodes:
sudo yum install pcs pacemaker fence-agents-all sbd lvm2-cluster gfs2-utils sudo systemctl disable firewalld sudo systemctl stop firewalld echo 'manager' |sudo passwd hacluster --stdin sudo systemctl start pcsd.service sudo systemctl enable pcsd.service
On one node:
sudo pcs cluster auth -u hacluster ha1.gbmdc.dc ha2.gbmdc.dc sudo pcs cluster setup --start --name spyderdb ha1.gbmdc.dc ha2.gbmdc.dc sudo pcs cluster enable --all
Configure fencing
Edit /etc/modules-load.d/softdog.conf on both nodes:
softdog
Execute:
sudo systemctl enable systemd-modules-load sudo systemctl start systemd-modules-load sudo pcs stonith sbd device setup --device=/dev/sdb sudo pcs cluster stop --all sudo pcs stonith sbd enable sudo pcs cluster start --all sudo pcs property set stonith-watchdog-timeout=10 sudo pcs stonith create sbd fence_sbd devices=/dev/sdb
Edit /etc/sysconfig/sbd on both nodes:
SBD_DEVICE="/dev/sdb"
Reboot nodes. After check if the output of this command is correct:
ps aux | grep -e COMMAND -e "sbd: watcher: /dev" | grep -v grep
Configure GFS2
sudo pcs property set no-quorum-policy=freeze sudo pcs resource create dlm ocf:pacemaker:controld op monitor interval=30s on-fail=fence clone interleave=true ordered=true
On both nodes:
sudo /sbin/lvmconf --enable-cluster sudo mkdir /spyderha
On one node:
sudo pcs resource create clvmd ocf:heartbeat:clvm op monitor interval=30s on-fail=fence clone interleave=true ordered=true sudo pcs constraint order start dlm-clone then clvmd-clone sudo pcs constraint colocation add clvmd-clone with dlm-clone sudo pvcreate /dev/sda sudo vgcreate -Ay -cy spyderDB_VG /dev/sda sudo lvcreate -L10G -n spyderDB_LV spyderDB_VG sudo mkfs.gfs2 -j2 -p lock_dlm -t spyderdb:gfs2db /dev/mapper/spyderDB_VG-spyderDB_LV sudo pcs resource create clusterfs Filesystem device="/dev/mapper/spyderDB_VG-spyderDB_LV" directory="/spyderha" fstype="gfs2" "options=noatime" op monitor interval=10s on-fail=fence clone interleave=true sudo pcs constraint order start clvmd-clone then clusterfs-clone sudo pcs constraint colocation add clusterfs-clone with clvmd-clone
Problems
Skipping clustered volume group spyderDB_VG
vgchange -cn $vgname --config 'global {locking_type = 0}'
Resource HTTPd example
Edit /etc/sysconfig/selinux:
SELINUX=permissive
On both nodes:
sudo setenforce 0 sudo yum -y install httpd sudo sed -i 's/\/var\/www/\/spyderha/g' /etc/httpd/conf/httpd.conf
On one node:
sudo mkdir /spyderha/cgi-bin sudo mkdir /spyderha/html sudo pcs resource defaults resource-stickiness=15000 sudo pcs resource create httpd systemd:httpd --group spyderdb sudo pcs resource create VirtualIP IPaddr2 ip=192.168.122.20 cidr_netmask=24 --group spyderdb sudo pcs constraint order clusterfs-clone then httpd
Disable monitoring of resources
pcs resource update httpd op monitor enabled=false
Monitor but do nothing
pcs resource update httpd op monitor on-fail=block