Cluster Server 8.0 Agent for EMC SRDF/Star Configuration Guide - Windows

Workload hosts and storage failure

In this scenario, the workload hosts and the storage array become unavailable. The original sync and async RDF links and the workload ICMP heartbeat links are lost. Under these conditions, you must use VCS to switch the application to either the sync target site or the async target site. While switching, you have a choice of using the data at either the sync or async target site. Set the UseMostRecentData and KeepDataOn attributes to specify which data you want to use.

If you switch to the sync target site, it becomes the new workload site. The agent then establishes a new async RDF link from this new workload site to the async target site.

At this point, no synchronous replication is available. Star is not in the PROTECTED state.

After the original workload site is available, you can regain Star protection by making the original workload site the new sync target site. You can then use the agent entry points to transition Star to the PROTECTED state.

See EMC SRDF/Star agent functions.

If you switch to the async target site, it becomes the new workload site. The agent then establishes a new async RDF link from this new workload site to one of the remote sites as specified by the asyncSite attribute.

At this point, no synchronous replication link is available. Star is not in the PROTECTED state.

Star protection is not available until the original workload site is recovered.