CPS and Exchange Logs not flushing on full backup

[Jason_Novak]

Running CPS for exchange with a full and flush scheduled for Saturday night. flushing of the logs failed with the following error V-79-57344-65193. Since logs didn't get flushed they are now continually growing any input is appreciated. Error code talks about traversing mount points and symantec is not going to fix till next major release?????

[perry_baker]

If something happens and a CPS Recovery Point can't truncate the logs then a full backup will have to be run.

If in your case the actual Full backup cannot truncate the logs then stop using this job while you try to determine what is going wrong.

Create a new Exchange Full backup to disk as is the BE/CPS/Exch backup and see if it will truncate the logs; you can see if the logs were truncated in the application log on the exchange server.

If that works start with a new continuous exchange backup from the beginning and monitor it.

If that failed then go back and start with a normal backup of the IS to tape and again verify if transaction logs were truncated on the exchange server.

If none of the backups will truncate the logs then try using NTBackup on the Exchange server itself and see if it can truncate the logs. If NTBackup can't do it then you would want to open a case with MS to determine what is wrong. If NTBackup is successful, then try BE again...if BE still fails then you should open a support case.

[Jason_Novak]

Perry,

Thank you for your response. I am able to flush the logs on both Windows Exchange backups and BE GRT backups to disk. The backup functionality of the exchange server is working properly. I did delete the cps backup job for exchange and recreated it and monitored it for the day. What I noticed is that it was copying the logs from the exchange server to the CPS server but not removing them. I had a restore point configured for every hour. Am I right in assuming that as the log is copy to the CPS server it is removed on the exchange server? If this is the case this wasn't happening. I did not let it continue because I didn't want my log files filling up the disk waiting for Friday's full to flush them.

I did open a case with you guys and the fix was to wait for the next service pack to be released in 2-4 weeks?????

Jason

[perry_baker]

The logs don't get truncated on the Exchange server until the Recovery Point runs.

What happens is the Recovery Point runs and then the BE media server contacts the Remote Agent on the Exchange server and tells it to tell Exchange to truncate the logs up to that point in time.

All CPS does is replicate the logs so that BE can organize them with the previous Full backup to generate a pseudo Full backup inclusive of the current logs.

So if the Recovery Point runs hourly then the logs should get truncated at those times.

You should be able to look at the time the Recovery Point ran successfully then go to the Exchange server, look in the Application Log and examine the entires from Exchange. It should show that a backup started at the time the Recovery Point ran, another entry will show the range of logs backed up and yet another entry should show the logs were truncated.

I have run this in my lab on a regular basis and it does work. The only time I had a problem was when my CMS was shut down for a day and the Recovery Point failed, once my CMS was running again I re-ran my Exchange Full backup and my Recovery Point backups started to function again.

I'm unaware of a problem related to that process that we are hotfixing. I apologize for my ignorance if there is one, I will have to investigate this possibility and will look to re-post to this thread.

[Jason_Novak]

Thanks for the clarification. I will give another go and see what happens. Will keep you posted.

[Jason_Novak]

Perry,

Does the first full backup need to run in order to start getting the logs to flush? Exchange CPS has beed running for about 2 1/2 hours now with recovery points at every hour and no logs have been flushed yet. See log below from CPS console. I have not seen anything in the BE Console.

Jason

Job Log for 'XXXX-EXCH-BE - Backup Exec'
Date Created: 3/7/2007 3:48 PM

3/7/2007 3:03 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' is running with no errors.

3/7/2007 3:03 PM Eastern Standard Time - The pair on source computer 'XXXX-EXCH-BE', destination computer 'XXXX-BACKUP' has started.

3/7/2007 3:03 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' is starting all source and destination pairs.

3/7/2007 3:03 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' has stopped.

3/7/2007 3:03 PM Eastern Standard Time - The pair on source computer 'XXXX-EXCH-BE', destination computer 'XXXX-BACKUP' for job 'XXXX-EXCH-BE - Backup Exec' has completed.

3/7/2007 3:03 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' is stopping.

3/7/2007 1:54 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' is running with no errors.

3/7/2007 1:54 PM Eastern Standard Time - The pair on source computer 'XXXX-EXCH-BE', destination computer 'XXXX-BACKUP' has started.

3/7/2007 1:54 PM Eastern Standard Time - Job 'XXXX-EXCH-BE - Backup Exec' is starting all source and destination pairs.

End of log

[perry_baker]

Yes you do need the first Exchange Full to run and truncate the logs initially. Then, with each recovery point the logs should get truncated in the manner I described.