Link to home
Start Free TrialLog in
Avatar of piotrmikula108
piotrmikula108Flag for United States of America

asked on

URGENT - Exchange accumulated large amount of transaction logs - how to get rid of them

Running SBS2003 with Exchange

100GB of t-logs and disk is running out of space

We do image backup of our entire server so backup doesn't clear T-logs. Is it safe to just delete them or switch to circular logging?

Thank you
ASKER CERTIFIED SOLUTION
Avatar of Jamie McKillop
Jamie McKillop
Flag of Canada image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of piotrmikula108

ASKER

We have a really good backup solution, image snapshots taken every 15min.

I will check your recommendation.

How about going forward - to avoid this situation in the future - is circular logging a good approach?
Even taking snapshots every 15mins, you will lose data in the event of a failure. You will just lose 15mins of data. If you are comfortable with that, you can enable circular logging.

JJ
recovery up to the minute is realative though.  if the server fails and you do not have the logs anywhere you are not going to be able to restore to anything except your last "valid" backup point.  But that is not really the basis if my comment..  lol

Question I have is this, your snapshot backup.  Is it a solution similar to DPM where it is actually syncing the logs and doing a DB on schedule?  Or is it actaully snapping the entire DB every 15 min?

Thanks,
Tyson
SOLUTION
Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
it's Zenith's BDR solution - uses StoargeCraft ShadowProtect backup solution - it backs up the entire server - so should the DB crash we only loose about 15 min, which is acceptable in our enviroment - and then we can restore the entire DB
That's why you place the database and transaction logs on separate LUNs. If you lose one LUN or the other, you will be able to recover up to the minute of failure. The only way you lose any data is the unlikely event both LUNs die at the same time or your whole server melts in a fire.

You need to to do Exchange aware backups to properly protect and ensure the smooth operation of Exchange.

JJ
Then based on your comment.  if you are ok with loosing up to 15min of data you could indeed use circular logging.  But again just as a reminder if it fails you COULD LOSE UP TO 15MIN OF DATA.

You could also use a script to clear out the log files that are already committed but that gets a little interesting and playback is not always successful in that instance anyway.
Ok - we are Zenith resellers, so I'm familiar with your system.  You need to get in touch with your Zenith support people and get them to configure your backups  on the Exchange server so that the last backup of the day (usually at 11:30PM) is a VSS backup - or do this yourself if you are managing your Zenith system.  This will ensure that your 11:30 Exchange backup is a full online backup and will purge the logs automatically. If you're managing the system but not familiar with how to adjust this setup, I can probably help walk you through it, or you could get Zenith to do an on-line chat with you to get it done.
hypercat, hello

are you talking about BDR's Clean State backup?
No, although that would probably work, I'm not sure. That's not the way I configure it though. There are settings in the StorageProtect software on the BDR that specify that the last backup of the day should use VSS.  Because Exchange is a VSS-aware application, this would allow the StorageProtect backup to act as a full on-line database backup and purge the logs. That's how we always do it.
ok, can you send more detailed instructions? is it somewhere on the official http://www.saaztraining.com/ website?
Yes - it's in the BDR Administration Guide (the version I have is dated 1/20/2010), on page 109. These are the instructions on how to edit a backup job. What you want to do is edit the portion of the backup job that schedules VSS jobs so that a VSS job is scheduled at least once a day - the schedule dialog box is shown in the graphic on p. 109. If you have any questions after taking a look at the instructions, post them here.