Import and Deport task of a Cluster Diskgroup fails due to timeout issues caused by VxVDS plugin

Problem

 VxVDS refresh operation interferes with disk group import/deport operation resulting in timeout and delay. This happens as both refresh and import/deport disk group processes try to read disk information at the same time.
 

Error Message

 The Cluster Diskgroups will fail to import/ deport due to timeout errors

Cause

VxVDS.exe is a VDS Dynamic software provider that manages SFW dynamic disks and volumes in the world of VDS.  It is a VEA client to Storage Agent and it listens to VEA events sent by Storage Agent whenever there are updates and activities on SFW objects.  It use that to update VDS on changes to SFW disks and volumes.Onlining and offlining a VCS resource group involves quite a few operations on Storage Agent and these operations triggers many event notifications which VxVDS.exe is listening to.  And it takes VxVDS.exe time to act on these events.

VxVDS.exe needs to query Storage Agent for changes in disk and volume objects and update VDS objects in many of those events , this includes a failover or move of the Volume manager Disk resources in a VCS or MSCS or WFC.

In general, the performance of VxVDS.exe should not be an issue. The timeout issue described in this article has only been seeing with large LUN/Disk configurations of over 100 disk. Each path to each disk counts on the total number of devices presented to host.

Solution

The issue is fixed by the following changes in Storage Foundation for Windows. To abort the refresh operation if disk group import/deport operation is in progress. Instead of doing full refresh on all the disk groups, refresh operation will now be performed only on disk groups that are being imported/deported. This is resolved via a hotfix. The hotfix is available for download from the following link :

https://sort.Veritas.com/patch/detail/5005

Applies To

SFW-HA 5.1 SP2

SFW-with Windows Cluster Option 5.1 SP2