Replication

To protect client data that was backed up to IBM Spectrum Protect servers, and to prevent system downtime when a server is unresponsive, replicate the client data to secondary servers.

About this task

Replication is the process of incrementally copying client data from one server to another to enable reliable failover and failback processing. The server to which the client data was originally backed up is called the source replication server. Replication processing copies the client data from the source replication server to a secondary server called the target replication server.

After you identify the replication needs of your environment and determine the replication configuration that meets those needs, use the Operations Center to configure replication processing. To configure replication, use the Operations Center to specify the source-to-target server pairs, the data to be replicated, and the time and frequency of replication processing.

After replication is running at scheduled intervals, you can then use the Operations Center to check replication status. The goal of replication is to maintain the same level of files on both the source and target servers. By using the Operations Center, you can verify that the replication processes completed successfully and that replication performance provides timely coverage. If there is a replication error or performance shortcoming that affects data currency, you can address the problem.

For example, if replication performance is slow and, as a result, the client data on the target server is not current enough to satisfy failover and failback requirements, you might change your replication configuration or the associated hardware. Because replication currency relies on sufficient, dedicated hardware resources, you might respond by increasing your hardware and network capabilities. You might also solve the problem by changing your replication configuration to make better use of your existing hardware and network resources.

If scheduled replication processing fails, you can use the Operation Center to start replication processing after the cause of the failure is resolved. For example, if a network outage caused scheduled replication processing to fail, you can start replication processing after the network is restored.

Two ways to configure replication are available. In versions earlier than IBM Spectrum Protect 8.1.13, the only way to configure and start replication was by using command-initiated replication. With command-initiated replication, you define a replication server pair by using the SET REPLSERVER command, and create administrative schedules for the REPLICATE NODE and PROTECT STGPOOL commands. Although you do not issue these commands directly in the Operations Center, certain actions and wizards in the Operations Center issue the commands in the background or define administrative schedules for the commands.

Starting in version 8.1.13, you can use storage-rule-initiated replication. With storage-rule-initiated replication, you configure and schedule replication by defining replication storage rules.

Consider using replication storage rules: If you are currently using command-initiated replication, consider using storage-rule-initiated replication instead. To review the high-level steps for moving to replication storage rules, see Transitioning from command-initiated replication to storage-rule-initiated replication. Replication storage rules are designed to have certain advantages over command-initiated replication. For example, by defining multiple replication storage rules, you can configure replication of client data from the same source server to more than one target server. Another advantage of using replication storage rules is that you don't issue the PROTECT STGPOOL command separately. Replication storage rules combine the functionality of the REPLICATE NODE and PROTECT STGPOOL commands into one replication operation.