Vmware vsphere metro storage cluster configuration – HP 3PAR Operating System Software User Manual

Page 142

Advertising
background image

in a peer-persistence configuration, additionally configured for automatic failover, to post updates
regarding their operational presence and health.

Storage systems also poll the quorum witness for the current operational status of their partner
storage systems within each distinct peer-persistence pairing, conceptually referred to as a quorum.
The quorum witness and related infrastructure provide the ability to differentiate a simple RC link
failure from a source system failure, through detection of a coincident cessation of updates from
the source storage system within the quorum, thus providing a reliable trigger for automatic
peer-persistence failover.

For correct operation, and to avoid unnecessary automatic failovers, the quorum witness should
be located on a third site, where events that might impact site 1 or site 2 cannot impact the quorum
witness site at the same time. In addition, the quorum witness is connected to arrays on both sites
using non-remote-copy links, so that issues with remote-copy links cannot impact connectivity with
the quorum witness.

With the above two configuration characteristics (site and link independence), quorum witness
helps determine the nature of the failure:

Are the two arrays alive, but not communicating because of a link failure? the quorum witness
is still receiving and relaying updates from both of the arrays—no failover required.

Has one of the arrays/sites failed? The quorum witness would not be receiving updates from
one of the arrays or sites that has failed, and the other array or site would be aware that
updates were lacking—a failover is required.

VMware vSphere Metro Storage Cluster Configuration

There are two types of VMware vSphere Metro Storage Cluster (vMSC) configuration, depending
on how the hosts are connected to the storage arrays. With HP 3PAR OS 3.1.2 MU2, HP supports
vMSC uniform configurations.

Uniform host access (see

Figure 52 (page 143)

) is a VMware-stretched cluster configuration where

each host can access storage resources available at both the sites. It is geared towards protecting
against a storage failure at a site.

142

HP 3PAR Peer Persistence with Transparent Failover

Advertising