Permanently Removing a Solaris Cluster 3.2 or 3.3 Node From Cluster

1
3/17/13 Document Display https://support.oracle.com/epmos/faces/ui/km/SearchDocDisplay.jspx?_adf.ctrl-state=k3z1s1nqr_96 1/1 Permanently Removing a Solaris Cluster 3.2 or 3.3 node from cluster [ID 1008264.1] Modified: Aug 4, 2012 Type: HOWTO Migrated ID: 211324 Status: PUBLISHED Priority: 3 Applies to: Solaris Cluster - Version 3.2 12/06 to 3.3U1 5/11 [Release 3.2 to 3.3] All Platforms Goal In the case of a severe node failure, it may become necessary to remove one node that can no longer join the cluster from the existing configuration, repair the node or reinstall from scratch, and add it (back) to the cluster. Fix From one of the remaining nodes in the cluster perform the following: 1. remove the failed node from the Nodelist property of all resource groups and device groups that the failed node could have mastered for resource groups, this command is # clrg remove-node -n <failednodename> <resourcegroupname> for device groups this command is # cldg remove-node -n <failednodename> <devicegroupname> 2. remove the failed node from quorum a. # clq disable <failednodename> b. on 2 node clusters # cluster set -p installmode=enabled c. remove the quorum device # clq remove qdname 3. clear the node from the node still in the cluster # clnode remove <failednodename> # clnode clear -F <failednodename> This will remove the node from the cluster configuration, for the participating nodes.

description

Permanently Removing a Solaris Cluster 3.2 or 3.3 Node From Cluster

Transcript of Permanently Removing a Solaris Cluster 3.2 or 3.3 Node From Cluster

Page 1: Permanently Removing a Solaris Cluster 3.2 or 3.3 Node From Cluster

3/17/13 Document Display

https://support.oracle.com/epmos/faces/ui/km/SearchDocDisplay.jspx?_adf.ctrl-state=k3z1s1nqr_96 1/1

Permanently Removing a Solaris Cluster 3.2 or 3.3 node from cluster [ID 1008264.1]

Modified: Aug 4, 2012 Type: HOWTO Migrated ID: 211324 Status: PUBLISHED Priority: 3

Applies to:

Solaris Cluster - Version 3.2 12/06 to 3.3U1 5/11 [Release 3.2 to 3.3]All Platforms

Goal

In the case of a severe node failure, it may become necessary to remove one node that can no longer join thecluster from the existing configuration, repair the node or reinstall from scratch, and add it (back) to the cluster.

Fix

From one of the remaining nodes in the cluster perform the following:

1. remove the failed node from the Nodelist property of all resource groups and device groups that the failed nodecould have mastered for resource groups, this command is

# clrg remove-node -n <failednodename> <resourcegroupname>

for device groups this command is

# cldg remove-node -n <failednodename> <devicegroupname>

2. remove the failed node from quorum

a. # clq disable <failednodename>b. on 2 node clusters # cluster set -p installmode=enabled c. remove the quorum device # clq remove qdname

3. clear the node from the node still in the cluster

# clnode remove <failednodename>

# clnode clear -F <failednodename>

This will remove the node from the cluster configuration, for the participating nodes.