Backup failures are very common in SCCM and I’ve NOT seen much of posts in this topic except 2 or 3. Recently, I’ve stumbled with a backup issue. SMS backup was getting failed, restarted the services, server, checked SQL connectivity, applied hotfixes etc… however no luck. At last, the post published by Rob Marshall – MVP helped to do further dig in. As mentioned in his post, my primary server also a had very bad backlog and this was because of huge number of secondary servers and communication issues with them. So the initial issue was, obviously, with registry key “\SMS_SCHEDULER: Routing Packages”. To clear out that mess, I followed the steps mentioned in Rob’s post.
So all went well after that ? Was the backup successfully completed? NO !! Why ?
So again, I stumbled with following errors in smsbkup.log, “After DoSnapshotSet SQL Writer status = FAILED_AT_FREEZE” and “Error: VSS_E_WRITERERROR_NONRETRYABLE. Error Code = 0x80,042,3f4“. Now what ? Back to basics, in SMS 2003 days, during SMS backup all the other SMS related services used to be in stopped state. This is changed in SCCM 2007 with the integration of VSS. So, in SCCM 2007 and 2012, the SMS services will be down/stopped/restarted for a fraction of second during the start of SMS backup process.
In my Scenario, SMS backup was not able to stop the “SMS Executive” service. So, where did I get the clue about this? Again ConfigMgr Admins best friend, the log files. In my case, it was smswriter.log. In smswriter.log, I was able to see that the SMS Component Manager service was stopped successfully and the next step that was supposed to happen was stopping the ‘SMS Executive’ service. But the backup was not able to stop the service and it was getting timed out. Now, as a workaround, I’ve manually killed smsexec.exe (exe for SMS Executive Service) from task manager while it was in stopping state. That is it !! Wow!! This resolved my backup issue. The backup was successful.
So what was the cause of the issue? Why the backup process was not able to kill or stop the SMS Executive Service (smsexec.exe) ? Yes, if you noticed above, I had already mentioned that primary server had some communication problem with secondary sites (some of them) and primary server had a very bad backlog. The problem was SMS Executive service was very busy in processing those requests and hence the backup service was not able to stop SMS Executive Service.