Table of ContentsLibraryView in Frames

How the site recovery plug-in works with MetroCluster

OnCommand Site Recovery Plug-in for Microsoft enables you to quickly recover from various storage system and site-wide failures by taking different actions depending on the type of failure the MetroCluster cluster resource detects.

In general, the site recovery plug-in works for a MetroCluster storage environment by taking the following actions:
  1. The plug-in verifies that the storage used by the application that you have added a MetroCluster cluster resource to is online.
  2. The plug-in allows the application to start.
  3. When the plug-in detects a storage disruption, it initiates a recovery scenario.

Recovery from storage system failures

Storage system failures involve a complete loss of the storage controller and disk shelves, resulting in the following recovery scenario:
  1. The site recovery plug-in detects the storage failure.
  2. The WSFC brings the application offline and restarts the plug-in.
  3. The plug-in verifies the disaster conditions.
  4. The plug-in initiates "cluster failover on disaster" to restore system availability at the disaster site.
  5. The plug-in monitors the storage system for takeover completion.
  6. The WSFC restarts the application.

Recovery from site-wide failures

With site-wide failures, there is a complete loss of storage system and Windows cluster nodes, resulting in the following recovery scenario:
  1. The WSFC moves application and site recovery plug-in instances to the surviving cluster node at the surviving site.
  2. The WSFC restarts the plug-in.
  3. The plug-in verifies the disaster conditions.
  4. The plug-in initiates cluster failover to restore availability at the disaster site.
  5. The plug-in monitors the storage system to be sure takeover has completed.
  6. The WSFC restarts the application, which is now using storage at the surviving site.

Recovery from other failures

Other failures include the following:
  • Controller and shelf failures, which are managed by MetroCluster
  • Windows cluster node failures, which are managed by the WSFC
  • Network failures, which are handled by the WSFC as long as a quorum can be established (which depends on the quorum settings)