SUMMARY
When a backup fails, it could be due to timeout issues. The timeout values need to be increased in the master.ini on the server in question in order to stop this from happening.
ISSUE
Backups fail or network timeout issues are experienced. One of the following errors is encountered:
Error
WARNING: Network connection stopped (time out).
FATAL: No backup data received from client.
Error
Error: VSS_E_WRITERERROR_TIMEOUT
RESOLUTION
On the client server having the backup errors, go to the directory: C:/PCBP/MASTER.ini file, open MASTER.ini as an administrator, and make the following changes:
-
Add the following on the line to the section [BProfessional]
NetTCPKeepAliveTimeout=120 ; added for troubleshooting
-
Change to following values in the section [usnap]. You want to increase the value by 600 of the value that it is already set to: NOTE: all 3 values should be increased by the same amount. Time is in seconds. increasing to beyond 20 minutes has no value as Microsoft's own internal timeouts cannot be adjusted further.
[usnap] USnapEnable=TRUE USnapKeepalive=1800 ; increased value by 600 USnapMutexTimeout=1420 ; increased value by 600 UsnapVssAsyncFinished=1200 ; increased value by 600
-
Optionally, add to following values (in red) in the section [usnap] section of the master.ini (if not already added) NOTE: This is only necessary if vssadmin list providers returns more than 1 VSS provider, which may be caused by some 3rd party backup products or SAN vendor driver integration.
[usnap] USnapProvider=System ; added force use of the Provider type: System USnapEnable=TRUE
-
Save the MASTER.ini file.
-
Restart the BP_Agent service to make sure these changes are registered.
-
Kick off a new backup job and verified it has successfully completed.
CAUSE
Failed backups could be caused by timeout issues related to VSS processes. The timeout values need to be increased in the master.ini on the server in question in order to temporarily mitigate this issue.
Note, VSS operations should normally complete in 30-90 seconds. Unitrends default tim-out is 10 minutes for these processes. For servers taking longer than 10 minutes that successfully complete in less than 20 minutes, this indicates a SEVERE performance or configuration issue with the client server that shod be addressed and resolved by a Microsoft expert. Unitrends considers adjusting this timeout to be a temporary measure in lieu of immediate server repair or rebuild. Customers are advised to not simply accept the adjustment of these values as a fix, but are instead encouraged to correct the issues in their server that are resulting in extended VSS query times.
Even after adjustments to these settings, intermitant VSS timeouts may still occur. If VSS has failed a reboot of the service and/or server may be required to resolve, and the issue may return.
VSS is a critical Microsoft component all backup vendors leverage that the backup vendor does not have direct control of. Unitrends cannot directly troubleshoot VSS related issues, but VSS related issues can not only prevent successful backup, but can exacerbate other server issues possible causing undue load or production hang/crash if left unchecked. Unitrends uses these timeouts to highlight when a server is under strain and when it should be reviewed by a Microsoft professional.