1- Make sure all packages run on primary node. In this case you want to remove clnode40 from cluster.
clnode40/root/home/root#cmviewcl
CLUSTER STATUS
ssdd_cluster up
NODE STATUS STATE
clnode40 up running
clnode53 up running
PACKAGE STATUS STATE AUTO_RUN NODE
DBssdd up running enabled clnode53
NFSssdd up running enabled clnode53
CIssdd up running enabled clnode53
rsyncssdd up running enabled clnode53
2- Halt secondary node. I'm halting the node from inside of the node itself, but I'm suggesting to halt it from the primary node. Because later you need to remove the node and this can't be done from the node to be removed.
clnode40/etc/cmcluster/ssdd#cmhaltnode -f clnode40
Disabling all packages from starting on nodes to be halted.
Warning: Do not modify or enable packages until the halt operation is completed.
Waiting for nodes to halt ..... done
Successfully halted all nodes specified.
Halt operation complete.
clnode40/etc/cmcluster/ssdd#cmviewcl
CLUSTER STATUS
ssdd_cluster up
NODE STATUS STATE
clnode40 down halted
clnode53 up running
PACKAGE STATUS STATE AUTO_RUN NODE
DBssdd up running enabled clnode53
NFSssdd up running enabled clnode53
CIssdd up running enabled clnode53
rsyncssdd up running enabled clnode53
3- pull current cluster/packages config, these command will save the config in the file name you specified.
cmgetconf -p rsyncssdd rsyncssdd.conf
cmgetconf -c ssdd _cluster ssdd _cluster.conf
4- Remove the reference line for the departing node. The reference line begins with the string:
NODE_NAME (in cluster config)
or
node_name (in package config)
5- If removing a node from the cluster results in a one-node cluster, remove all lines containing the references that follow from the cluster ASCII file:
FIRST_CLUSTER_LOCK_VG
FIRST_CLUSTER_LOCK_PV
The lock disk function is not used in a one-node cluster.
6- apply the new modified config
clnode53/etc/cmcluster/ssdd#cmapplyconf -C ssdd _cluster.conf -P DBssdd.conf -P NFSssdd.conf -P CIssdd.DBssdd.conf -P NFSssdd.conf -P CIssdd.conf -P rsyncssdd.conf
MAX_CONFIGURED_PACKAGES configured to 300.
MAX_CONFIGURED_PACKAGES configured to 300.
NFSssdd.conf:490: service_halt_timeout value of 0 is equivalent to 1 sec.
Modifying the cluster locking mechanism from lvm to majority while cluster ssdd_cluster is running.
Deleting FIRST_CLUSTER_LOCK_PV /dev/disk/disk98 from node clnode53 while cluster is running.
Removing configuration from node clnode40
Modifying configuration on node clnode53
Deleting node clnode40 from cluster ssdd_cluster
Modify the cluster configuration ([y]/n)? y
Completed the cluster creation
7- verify node is not visible in cmviewcl
clnode53:/etc/cmcluster/ssdd# cmviewcl
CLUSTER STATUS
ssdd_cluster up
NODE STATUS STATE
clnode53 up running
PACKAGE STATUS STATE AUTO_RUN NODE
DBssdd up running enabled clnode53
NFSssdd up running enabled clnode53
CIssdd up running enabled clnode53
rsyncssdd up running enabled clnode53
8- Set AUTOSTART_CMCLD=0 in /etc/rc.config.d/cmcluster. Set AUTO_VG_ACTIVATE=1 in /etc/lvmrc.
Command outputs are from me, cluster commands are lifted from HP site: http://support.hp.com/us-en/document/c01058926
#WHATYEARISTHIS??
No comments:
Post a Comment