Serviceguard Extended Distance Cluster (EDC) with VxVM/CVM Mirroring on HP-UX, May 2008

Summary of Best Practices

This section summarizes the best practices for EDC with VxVM/CVM which have been discussed in

this document - it can be used as a checklist. For information about why and how these practices are

recommended, the entire document should be read.

General best practices, which also apply for non VxVM/CVM EDC:

• Use a common redundant link for heartbeat and storage traffic.

• Use two fault tolerant DWDM, CWDM, SONET, or SDH boxes per site instead of one.

• Protect the quorum service in a separate cluster.

• Configure cluster parameters and I/O timeout values to ensure that cluster reformation can complete

before I/O timeout expires.

VxVM/CVM specific best practices for EDC:

• Switch off hot relocation for all shared volumes in the EDC

• Configure enclosure-based naming

• Customize enclosure names to include the site name

• Mirror all volumes across enclosures (or sites, with VxVM/CVM 5.0)

• Always specify storage attributes like mirror=enclr enclr:DC1-XP12k1 enclr:DC2-

EVA8k1 when creating or changing volumes

• Configure mirrored DCO volumes for DRL an FMR

• When accessing a shared volume from only a single node, consider setting the read policy for that

volume in the package control script, so it prefers the local plex to enable local reads from the site

where the application package is running

• Consider monitoring the syslog for FC link, and mirror/plex failure messages to start the manual

resynchronization soon after links become available again

VxVM/CVM 5.0 specific best practices for EDC:

• Configure sites for all cluster nodes and storage devices

• Configure site-consistent disk groups and volumes

• Mirror across sites and always specify storage attributes like mirror=site site:DC1

site:DC2 when creating or changing volumes