During an event where a shutdown of the Hitachi Metrocluster is neccessary (ie. migration of storage cluster, power outage event where backup power is not available) a graceful shutdown of the Cluster is necessary in order to avoid a Raid group rebuild which may result in a cluster running in a degraded state for an extended period of time.
So here is the proper procedure for safe shutdown:
- Stop application and shutdown app servers
- Unmount all file systems
- Shutdown HNAS
- Split Pair
- Shutdown Storage Array
The split pair operation is crucial in order to avoid a rebuild of your raid group when systems come back online. During an event of an abrupt power loss, the AMS/Sync DR will not be able to determine the state of your raidgroup thus an automatic rebuild will be triggered which may cause the system to run in a degraded state for 10+ hours depending on the size of your raid groups.
It is advisable to perform the split pair operation using the Hitachi Storage Navigator Modular 2 (SNM2)
You need to power back the environment in reverse sequence:
- Power up Storage array
- Resync Pair
- Power up HNAS
- Mount all file systems
- Power up app servers and start the applications.
When bringing the cluster back up, the operations are reversed and the resync pair operation is necessary for Raid groups to recognize each other.