Our MDBDATA log folder has gotten full and it's causing users outlook to lose connection. To remedy the problem my co-worker suggested that we move old logs onto our D: drive since it's not being used. Now the D: drive is starting to get full. I've tried to run the MDBDATA backup but after awhile it fails. I'm not the best when it comes to Exchange server but could it be our old logs are moved onto another HDD? Any suggestions on what I can do?
2 Answers
You have two choices:
Perform a full backup using an Exchange-aware backup utility (like NTBackup). This will flush the transaction logs.
Enable circular logging on the storage group.
It's unclear to me why you can't get a backup to complete. More to the point, it's unclear to me that you're even bothering with backups. That's scary.
If you're not going to perform regular backups with an Exchange-aware backup program and/or the database (EDB / STM files) are on the same volume as the transaction logs then you might as well just switch into circular logging mode. You lose the ability to do point-in-time recovery and incremental/differential backups but it sounds like you don't particularly care about that anyway.
You're one database corruption away from losing all your Exchange data if you're not performing proper backups.

- 141,881
- 20
- 196
- 331
-
After running an NTBackup after an hour or so I get an error "Error: An inconsistency wasa enountered in the requested backupfile." It's also unclear to me why it's not work (My reason for seeking help). – stick_tech Jan 09 '13 at 20:03
If you are willing to take the risk, or you simply don't care about the old logs, you can delete them (up to say the last few days), and that would get you back in operation after an Information Store services restart. I don't recommend it and I'm sure it's highly against MS' best practice, but I've done it myself twice over the years and then forced a full backup that night of Exchange. It'll have to be a full backup and not a differential or incremental at that point.
See here for more info: http://support.microsoft.com/kb/240145
And here for a great walkthrough: http://www.msexchange.org/articles-tutorials/exchange-server-2003/planning-architecture/Exchange-log-disk-full.html

- 32,627
- 26
- 132
- 191
-
1Don't delete the logs after the last checkpoint file. That can be a couple days, but I've seen it be worse than that in some bad Exchange designs. Make sure you check when the last checkpoint was before you start deleting log files and make sure you take a good backup after you delete log files. You could also just enable circular logging on the databases and dismount/remount the stores (make sure you understand impact to recovery options if you do this). – Rex Jan 09 '13 at 16:12
-
-1 - Woah there! Don't ever manually delete Exchange transaction logs. – Evan Anderson Jan 09 '13 at 17:09
-
@EvanAnderson - I never said it was "proper" just that it can work, like Rex is saying, but I agree with the downvote actually. I wouldn't recommend it normally but I've had to do it in 2 extreme cases where Execs said "do it, we ain't waiting for a backup" – TheCleaner Jan 09 '13 at 17:21