Requirements for automated client failover

Before you configure or use the client for automated client failover, the backup-archive client and IBM Spectrum® Protect server must meet several requirements.

Ensure that the client meets the following requirements for automated client failover:
  • The primary server, secondary server, and backup-archive client must be running IBM Spectrum Protect Verson 7.1, or a later version.
  • The primary and secondary servers must be set up for node replication.
  • The client node must be configured for node replication on the source replication server by using the REGISTER NODE REPLSTATE=ENABLED or UPDATE NODE REPLSTATE=ENABLED server commands.
  • By default, the client is enabled for automated client failover. However, if the usereplicationfailover no option is specified in the client options file, either change the value to yes, or remove the option.
  • Valid connection information for the secondary server must exist in the client options file. During normal operations, this information is automatically sent to the client from the primary server.
  • To save the secondary server connection information that is sent from the primary server, the client must have write access to the dsm.opt file on Windows clients, and the dsm.sys file on AIX®, Linux®, Mac OS X, and Oracle Solaris clients. If the client does not have write access to these files, the secondary server information is not saved to the client options file, and an error is added to the error log.
  • Non-root users cannot use the default location for the node replication table. You must specify a different location by adding the nrtablepath option to the dsm.sys file. For more information, see Nrtablepath.
  • The following processes must occur before the connection information for the secondary server is sent to the options file:
    • The client must be backed up to the source replication server at least one time.
    • The client node must be replicated to the target replication server at least one time.
  • Failover occurs for client nodes that are backed up with client-node proxy support when both the target and agent nodes are configured for replication to the target replication server. When the target node is explicitly replicated, the agent node is implicitly replicated to the target replication server as well, along with the proxy relationship.
    For example, Node_B is granted authority to perform client operations on behalf of Node_A with the following server command: 
    grant proxynode target=Node_A agent=Node_B
    If both nodes are configured for replication with the replstate=enabled option in the node definition, when Node_A is replicated, Node_B and the proxy relationship are replicated as well.