Admin guy deleted some transaction log files while information store is online! - Exchange 2007
I was at my client's yesterday. While working on something else, the admin I was working with decided he would start deleting transaction log files on his Exchange 2007 server. I halted him right away but some already went into the recycle bin. I had him restore the those transaction log files from the recycle bin. But now symantec backup is detecting corruption in the transaction log files, therefore, not purging the log files at the end of the backup. I've restarted the server and restarted Exchange information store service quite a few times, but the no purging of transaction log file situation is not changing. I read some suggestions here that it's possible to turn on circular logging to purge the files, restart IS, then turn circular logging back off. Is this a safe bet? I understand the procedure to run eseutil /mh, but what if I get a dirty shutdown result, which is most likely the case. thanks!
June 28th, 2011 10:08pm

If you did a normal shutdown / restart of the server, Exchange would have tried to commit all the outstanding log files before it dismounted the DB. As long as all the missing log files had been committed before they were deleted your DB should be OK. If not, I'd expect you'd know about it as soon as it tried to re-mount after the shutdown. I'd dismount the DB and run eseutil /mh and see if it it's in a clean shutdown state. I expect it will be. [string](0..33|%{[char][int](46+("686552495351636652556262185355647068516270555358646562655775 0645570").substring(($_*2),2))})-replace " "
Free Windows Admin Tool Kit Click here and download it now
June 28th, 2011 10:53pm

In addition with above, you can enable cirular logging for databse and later you can remove it, But here i don't think it will help for logical correuption. you can take one NT (streaming backup) backup of databse and then full backup thru symantc. When you take full backup exchange self correct issue of some log missing in Exchnage 2007. If still not wotj then you can go ahead with /mh and then /P (Last action)Anil MCC 2011,ITIL V3,MCSA 2003,MCTS 2010, My Blog : http://messagingschool.wordpress.com
June 28th, 2011 10:59pm

Anil, I don't see how a ntbackup would make a difference if there is log corruption. although I can have the guy try it and see what happens. I will let you know. Thanks.
Free Windows Admin Tool Kit Click here and download it now
June 29th, 2011 9:37am

When you're missing log files the only way to get your backups to purge is to either sucessfully recover those missing corrupted logfiles which usually ends up not be possible or you have to manually delete the commited log files yourself or by enabling circular logging. (It's usually recommended to just move the committed log files to a temp location instead of deleting as a fyi measure) Article is for 2003 but applies to 2007. How to remove Exchange Server transaction log files http://support.microsoft.com/kb/240145 James Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com
June 29th, 2011 10:02am

1. I've come across this a few times. Athough the IS is still online and the databases mounted, log file have been deleted, i.e the ones that have been written to the database. Admin decided to do this as during the day, log space was running out, option was not to distrupt service and make space avaliable, so logs file were deleted. (apthough they could have beem moved) 2. The issue you have here is that you cannot recover using thoses logs file and almost all backups fail until you complete a FULL backup. You must do a full backup if youre not doing this already and will sort it out. 3. Not usre if youre performing any other type of backup incre, diff etc..but do a FULL then revert back to your normal backups if they arent FULL. Sukh
Free Windows Admin Tool Kit Click here and download it now
June 29th, 2011 5:05pm

hi there, full backup is failing during integrity check of transaction log files in Symantec Backup Exec 2010 R2, that's why the logs aren't purging. So a full backup isn't purging the logs, that's why I need an alternative. I have recommended the admin to purge the logs by turning on circular logging, then restarting the microsoft exchange IS service as what others are recommending here. He will be doing this tonight, let's see what happens.
June 29th, 2011 5:21pm

Turning on circular logging may not reconcile the problem right away. The backup doesn't like that there are "holes" in the log file list (the file names should all be sequential, and they're not). Eventually new log files will overwrite the old ones, and all the log file names will all be sequential again, but it may take awhile, depending on how big the log file directory is, and how fast new log files are generated. Since that set of log files is not salvagable for recovery purposes, you could dismount the database and delete them (or move them to a another directory temporarily) and then re-mount the database, and the log files will restart at the next available sequence number, and the backups should be happy again. Just be careful to only get the log files, and leave the checkpoint file.[string](0..33|%{[char][int](46+("686552495351636652556262185355647068516270555358646562655775 0645570").substring(($_*2),2))})-replace " "
Free Windows Admin Tool Kit Click here and download it now
June 29th, 2011 5:51pm

I like the idea of moving out log files better. So leave only the .chk files and mailbox database.edb?
June 29th, 2011 10:24pm

Yes. The chk is the checkpoint file. It needs that to know what sequence number to use for the next log file.[string](0..33|%{[char][int](46+("686552495351636652556262185355647068516270555358646562655775 0645570").substring(($_*2),2))})-replace " "
Free Windows Admin Tool Kit Click here and download it now
June 29th, 2011 10:38pm

Problem resolved. I had the guy stop the IS service, moved the logs out, then restart the IS service. Apparently all logs have been committed and moving the logs out didn't pose a problem.
July 5th, 2011 2:25pm

Good to hear.[string](0..33|%{[char][int](46+("686552495351636652556262185355647068516270555358646562655775 0645570").substring(($_*2),2))})-replace " "
Free Windows Admin Tool Kit Click here and download it now
July 5th, 2011 2:35pm

On Tue, 5 Jul 2011 18:23:30 +0000, ronnieshih wrote: >Problem resolved. I had the guy stop the IS service, moved the logs out, then restart the IS service. Apparently all logs have been committed and moving the logs out didn't pose a problem. By removing the log files he created a new log file sequence (check the numbering of the new log files and they should be very low compared to the log files that were removed). That means that any previous backups are void. Make a new, complete, backup of that storage group *NOW* or you may lose a considerable amount of information if you need to restore the database. --- Rich Matheisen MCSE+I, Exchange MVP --- Rich Matheisen MCSE+I, Exchange MVP
July 5th, 2011 8:56pm

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics