-
Title
vRanger Backups going to a NVSD (NetVault SmartDisk) Repository randomly fail -
Description
When using an NVSD (NetVault SmartDisk) as a vRanger Repository for backups, backups randomly fail immediately after they begin with a "Timeout Expired" error. If running other repositories on the same vRanger install (e.g. CIFS shares, etc.), backups to those other Repositories do NOT exhibit the same problem.
-
Cause
This is usually seen on vRanger installs that use a remote (i.e. not local to the vRanger machine) SQL database install. The NVSD Repository/vRanger connection to it is quite sensitive to database latency, and the timeout period is quite short before vRanger aborts the backup task.
-
Resolution
1) Manually shrink the vRanger database as per the instructions below from Microsoft:
http://msdn.microsoft.com/en-us/library/ms189080(v=sql.105).aspx
2) Disable cataloging for the job (un-check the "Catalog" option in the job configuration). Cataloging savepoints resident on an NVSD Repository (especially for densely-populated (VMs with many levels of folders/files) induces a lot of latency to the backup process and seems to exacerbate the issue. If cataloging is deemed necessary, it will work, but could force a database shrink/cleanup of the catalog database more often.
3) If possible, move the vRanger database to the vRanger machine and use a local database install. A local install will not query the database across the network, and it appears not to have the same latency issues as a remote SQL install vis-a-vis the NVSD Repository.