Removing Data Nodes from an Active SSR Cluster
This procedure makes the following assumptions:
You are working with an active cluster and the SSR database is in production mode. If you are unsure what mode the database is in, see Viewing the SSR Database Mode (SRC CLI).
Your cluster currently contains four data nodes and you want to change to a two-data-node cluster. Data nodes can be added or removed only in pairs.
Changing the SSR cluster geometry (adding or removing data nodes) destroys all data in the SSR database. All data is lost. This procedure should be performed only during a maintenance window. Review this procedure in full before proceeding.
To remove two data nodes from an active cluster:
Reconfigure the cluster topology to a two-data-node geometry.
Commit the configuration.
See Committing a Configuration and Exiting Configuration Mode.
Place the database into maintenance mode.
See Placing the SSR Database into Maintenance Mode (SRC CLI).
Disable the component database on all management servers in the cluster (one by one).
Disable the component database on the two data nodes you removed from the cluster (one by one).
Enable the component database on all management servers in the cluster (one by one).
Restart the component database on the two remaining data nodes in the cluster (one by one).
Re-create the SSR database.
Restart all SSR client components.
Verify the cluster configuration changes by viewing the status of the database.
Place the database into production mode.
See Placing the SSR Database into Production Mode (SRC CLI).