Hello everyone! As a backup solution for a SBS 2003 we're doing daily disk images of all volumes; one full on the first day of the month and then differentials. These daily differential backups are getting bigger than they should be, only from Exchange databases partition. We have only 7 users and the Mailbox Store is about 9GB, very few spam and emails with attachments are unusual. On this server the default 'Maintenance Interval' for the Mailbox Store is set to run daily from 1:00AM to 5:00AM. Since the backups are sector oriented, I'm supposing that this 'Maintenance' is changing a lot of sectors and then daily images are growing larger than they apparently should be. Am I right? 'Maintenance Interval' could be disabled if a offline defragmentation was done on a monthly basis? (just before the full backup) Any help would be appreciated, thanks!

is the maintenance completing? check the app log. The maintenace also performs other tasks except reclaming whitr space so I wouldnt cancel/disable that. Make surr your backups are not intterupting the maintenance. If you wish you can perform offline defrav but this requires downtime for eseutil to run. Maybe create another DB and move users actoss & backup that DB. . Sukh

Hi Sukh828, thanks for the reply! Judging by event logs, maintenance seems to be completing; No, my backups are not interrupting the maintenance, they are scheduled for a different time. Do you run a daily backup of Exchange databases and they grow up more than expected too? Any more comments from you or anyone else would be welcome!

Have you confirmed maintenance is completing? I do full backups but they shouldnt grow at an alarming rate unless a mailbox is growing at an unexpected rate or something that causes this. I would keep a record of how much the DB is growing on a daily basis as well as each mbx in each DB. Make sure the maintenance is completing. Log as this i a spreadzheet which will help you. If needed create a nee DB and move users across and test again. Make sure it is the DB that is growing and no other files on that partition.Sukh

That's correct OLM will cause considerable block changes but given that you only have 7 users I doubt this would explain it if you're seeing larger images. In any case with only 7 users you don't need to run OLM daily you can actually set it for every 1-2 weeks on weekends. Running OLM too much can be detrimental because it can cause uncessary churn. James Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com

@Sukh828: yes, logs has no errors and are complete, maintenance is finishing successfully. @Jamestechman: Thanks, that's what I tought. Now it makes sense :) I would like to schedule this on a monthly basis, just before the full backup. The GUI doesn't allow me to do that. How could I run this as a scheduled task? Could the Maintenance run from command line?

I'm not aware might not be possible, what's typically done to extend it past a week is just to schedule very short intervals so it takes 1-2 weeks to complete a full pass or in your case a month. So in your case run it every saturday for 15-30mins than track it in your event logs to see how long it takes to complete a full pass. James Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com

Checking the logs, the last maintenance has taken only 16 minutes to complete. Last event was: The database "First Storage Group\Mailbox Store (SERVER)" has 168 megabytes of free space after online defragmentation has terminated. Would be really perfect to run maintenance from scheduled tasks. Only one pass, monthly, just before the full backup, making differentials as small as possible and keeping databases in perfect shape :) I know this is a tough question, but some method should exist to accomplish that :)

I have a hard time being convinced that it's OLM thats bloating your backup sizes then. It could be other things like AV, file level defragmentation etc. Are you sure no other IO activity is going on on that box?James Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com

I'm sure. As advised, AV is disabled for the Information Stores, which uses a exclusive partition. Defragmentation only runs scheduled, monthly and before the full backup. The first two differentials are tolerable in size, but now that the month is ending each one takes 10% of the whole database size, unjustifiable by the amount of mail that we see. What you said does make sense: 'Running OLM too much can be detrimental because it can cause uncessary churn'. To make sure, I could disable OLM temporarily, using custom schedule and deselect everything. But now I'm dreaming with that command line :D

Yup I would start off with disabling the OLM for 1-2 weeks. Why would you want to reduce your online maintenance window? Well, for one thing, you can increase your backup window (if you are performing backups on the active copy of a database). In addition, if you decrease your online maintenance window and you are performing VSS backups, this will reduce database churn and produce smaller snapshots or differential backups. http://blogs.technet.com/b/exchange/archive/2007/12/06/3404504.aspxJames Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com

Yup I would start off with disabling the OLM for 1-2 weeks. Why would you want to reduce your online maintenance window? Well, for one thing, you can increase your backup window (if you are performing backups on the active copy of a database). In addition, if you decrease your online maintenance window and you are performing VSS backups, this will reduce database churn and produce smaller snapshots or differential backups. http://blogs.technet.com/b/exchange/archive/2007/12/06/3404504.aspxJames Chong MCITP | EA | EMA; MCSE | M+, S+ Security+, Project+, ITIL msexchangetips.blogspot.com

I'll do that and report here. Thanks!