Matrix Co-Existence with VMware vCenter Site Recovery Management (SRM)

Technical white paper | Matrix Co-Existence with VMware vCenter Site Recovery Management (SRM)
2
Introduction
The Matrix Operating Environment 7.3 Update 2 release provides co-existence with VMware vCenter Site Recovery Manager
(SRM). This white paper documents the necessary steps to follow for the Matrix Operating Environment to co-exist with
SRM. Prior to using SRM and Matrix together, there are some minor configuration steps that must be done. In addition, there
are some procedures that must be followed so that Matrix can remain synchronized with the virtual machines during and
after SRM failover processing.
This document assumes that the reader is familiar with both Matrix and SRM operations. This is not a User Guide for how to
configure or use either product; consult the appropriate documentation for details.
Matrix documentation is available at www.hp.com/go/matrixoe/docs.
The VMware SRM 5.1 Documentation Center is available at https://pubs.vmware.com/srm-
51/index.jsp?topic=%2Fcom.vmware.ICbase%2FWelcome%2Fwelcome.html.
SRM high-level overview
VMware vCenter Site Recovery Manager (SRM) is an extension to vCenter that provides a disaster recovery capability for
virtual machines. There are two types of recovery operations (i.e. failovers) that can be implemented from the protected site
to the recovery site.
Planned Migration: The orderly evacuation of virtual machines from the protected site to the recovery site. Planned
Migration prevents data loss when migrating workloads in an orderly fashion. For Planned Migration to succeed, both sites
must be running and fully functioning.
Disaster Recovery: Similar to Planned Migration except that Disaster Recovery does not require that both sites be up and
running. During a Disaster Recovery operation, failure of operations on the protected site are reported but otherwise
ignored.
When virtual machines at the primary site (the protected site) are configured for protection, there is a placeholder shadow
VM created on the secondary site (the recovery site). Using Array-Based Replication (ABR), the storage is replicated from the
protected site to the recovery site by the storage subsystem and not by any software on the VMs themselves, the vmkernel,
or the Service Console. When a failover operation is done, the virtual machines at the protected site are shut down
(whenever possible) and copies of these virtual machines (placeholder VMs) at the recovery site will start up.
What happens during a failover operation
When you do a Recovery failover operation, a pre-defined Recovery Plan will be executed. The Recovery Plan defines the
complete set of steps needed to recover the protected VMs in one or more Protection Groups. At a high level, SRM will do
the following major steps:
1. Shut down the VMs on the protected site.
A. If you selected the SRM Planned Migration option, SRM will shut down the VMs in an orderly manner.
B. If you selected the SRM Disaster Recovery option, SRM will attempt to shut down the VMs on the protected site.
Depending on the reason for the Disaster Recovery failover (e.g. an ESXi host may be down), SRM may not be able
to properly shut down all of the VMs on the protected site. In the Disaster Recovery failover scenario, SRM will
ignore any errors while trying to shut down these VMs.
2. Synchronize the storage between both sites.
3. Stop the storage replication from the protected site to the recovery site and make it read only on the protected site.
4. Activate the storage and make it read/write on the recovery site.
Note
When using 3PAR arrays, the name of the datastores will change with each failover due to the way the 3PAR Remote
Copy functionality works. With each failover, the datastore names are changed to snap-xxxxxxxx-<original
datastore_name>.
5. Power up the VMs on the recovery site.