When you rename a cluster that is in a partnership, the partnership becomes invalid. You must fully unconfigure the existing partnership and create a new one that uses the new cluster name.
This procedure demonstrates how to rename one of the global clusters that is in a partnership. You can rename more than one of the clusters at the same time.
If the cluster that you rename belongs to more than one partnership, perform each step on all clusters that share a partnership with the cluster to rename, before you proceed to the next step in the procedure.
This task avoids production application downtime.
# geopg remove-resource-group application-resource-group protection-group
# geopg list protection-group
This task stops data replication.
# geopg stop protection-group -e global
# geopg delete protection-group
# geops leave-partnership partnership
# geoadm status
# geoadm stop
Verify that the geo-infrastructure, geo-clusterstate, and data-replication resource groups are deleted.
# clrg list # geoadm status
Follow cluster naming guidelines as described in Planning Required IP Addresses and Hostnames in Installing and Configuring the Disaster Recovery Framework for Oracle Solaris Cluster 4.4.
# cluster rename -c new-cluster-name
# cluster list
If clusters in the partnership are in different domains, include the domain in the /etc/hosts entry for each cluster.
# ping new-cluster There should be no response # echo "IP-address new-cluster" >> /etc/inet/hosts
# geoadm start
If the disaster recovery framework fails to start, and the failure is not due to problems with the new logical host, restart the common agent container on all nodes by using the cacaoadm restart command, then start the disaster recovery framework.
# geoadm status
# geops add-trust -c remote-partner-cluster[.domain-name]
# geops verify-trust -c remote-partner-cluster[.domain-name]
# geops create -c remote-partner-cluster[.domain-name] partnership
# geops join-partnership remote-partner-cluster[.domain-name] partnership
# geoadm status
Restarting the heartbeat initiates the heartbeat to read and store the new cluster name.
# svcadm disable svc:/system/cluster/gchb_resd:default # svcadm enable svc:/system/cluster/gchb_resd:default
This example renames the cluster newyork, in the paris-newyork-ps partnership, to chicago. The names of the nodes in this cluster are not changed, so phys-newyork-1 becomes a node in the newly named chicago cluster. The paris-newyork-ps partnership is first unconfigured. After the cluster is renamed, a new paris-chicago-ps partnership is created with the chicago cluster as primary and the paris cluster as secondary. The two clusters belong to the same domain, so the domain name is not specified to the commands.
phys-newyork-1# geopg remove-resource-group app-rg phys-newyork-1# geopg list examplepg phys-paris-1# geopg list examplepg phys-newyork-1# geopg stop examplepg -e global phys-newyork-1# geopg delete examplepg phys-paris-1# geopg delete examplepg phys-newyork-1# geops leave-partnership paris-newyork-ps phys-paris-1# geops leave-partnership paris-newyork-ps phys-newyork-1# geoadm stop phys-paris-1# geoadm stop phys-newyork-1# clrg list phys-newyork-1# geoadm status phys-paris-1# clrg list phys-paris-1# geoadm status phys-newyork-1# cluster rename -c chicago phys-newyork-1# cluster list phys-newyork-1# ping chicago phys-newyork-1# echo "192.168.10.1 chicago" >> /etc/hosts repeat on each node of the chicago cluster phys-paris-1# ping chicago phys-paris-1# echo "192.168.20.1 chicago" >> /etc/hosts repeat on each node of the paris cluster phys-newyork-1# geoadm start phys-paris-1# geoadm start phys-newyork-1# geoadm status phys-paris-1# geoadm status phys-newyork-1# geops add-trust -c paris phys-paris-1# geops add-trust -c chicago phys-newyork-1# geops verify-trust -c paris phys-paris-1# geops verify-trust -c chicago phys-newyork-1# geops create -c paris paris-chicago-ps phys-paris-1# geops join-partnership chicago paris-chicago-ps phys-newyork-1# geoadm status phys-paris-1# geoadm status phys-newyork-1# /etc/init.d/initgchb_resd stop phys-newyork-1# /etc/init.d/initgchb_resd start repeat on each node of the chicago cluster phys-paris-1# svcadm disable svc:/system/cluster/gchb_resd:default phys-paris-1# svcadm enable svc:/system/cluster/gchb_resd:default repeat on each node of the paris cluster
Perform the following tasks:
Create a new protection group and replicate it to partner.
Add device groups.
Add resource groups to the protection group and verify the configuration.
Follow procedures in the appropriate Oracle Solaris Cluster data-replication guide.