This section shows an example of an Oracle workload running across SRDF/Metro, on both sides, and demonstrates how it continues to run, despite a simulated disaster that causes the two storage systems to be partitioned (so that they can not communicate).
This figure shows the Metro state and the DR state of the protected storage group environments. It shows the Array IDs for the MetroR1, MetroR2 and the DR sites.
Figure 36. Running SRDF/Metro Smart DR Environment
Here you see the state of the SRDF/Metro Smart DR Environment is “green,” indicating that the system is functioning normally both in SRDF/Metro and DR.
Figure 37. Unisphere view of the SRDF/Metro Smart DR environment
Swingbench is a Java-based load generator for Oracle. Order Entry is one of the load-generating tests. It behaves like an OLTP application. This is a two-host cluster with 10 users per host for a total of 20 users.
Figure 38. Swingbench running Order Entry on two Host Cluster
Here is an example of tnsnames.ora that allows both load balance and failover.
(LOAD_BALANCE = ON)(FAILOVER = ON)
(ADDRESS = (PROTOCOL = TCP)(HOST = dsib1190)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = dsib1191)(PORT = 1521))
(SERVICE_NAME = slob)
When a failure occurs on the SRDF/MetroR1 side, this is what you can expect to see in Unisphere for PowerMax. Both Metro and DR status will indicate a failure.
Figure 39. Unisphere View of the Link failure on SRDF/MetroR1 side
Since cross-links are not used in this example, one of the two database instances in the cluster does not survive the failure. The user sessions are migrated to the surviving database instance and continue running. Swingbench shows that all the user sessions continue to run and none were lost.
Figure 40. Swingbench after link failure
The MetroR2 will become the new MetroR1 and its SRDF/A Leg will become active. The SRDF Groups for Array ID 449 remained online and the DR status is at “Warning” but not failed.
Figure 41. New MetroR1 with SRDF/A leg active
Figure 42. View after link failure
After the links are recovered, an Establish operation is needed to re-establish the SRDF/Metro relationship. Once it is done the Metro Status will no longer be failed. Repeat this step for all environments.
Figure 43. Re-Establish SRDF/Metro after links recovery
After the environment is re-established the new MetroR1 will remain. This does not impact the functionality of SRDF/Metro Smart DR.
Figure 44. Another view of the Re-Established Environment