Priv1.edb is 135GB in size - there's only 19 users

[gmono]

Hi
Can someone help me - I'm pretty new to exchange.
Priv1.edb is 135GB in size - there's only 19 users and not one mailbox is over 1 GB in size - something tells me this isn't right?! I tried an offline defrag but there wasn't enought space on the disk for it to finish?

Does anyone have any ideas?

Eachange 2003 - installed 2 weeks ago????

Thanks

 

[58sniper]

Look for event 1221 in the application event log. It should show you how much of that database size is white space.

Do you have mailbox limits in place?

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[gmono]

Hi

The database "First Storage Group\Public Folder Store (W2K3JOR001)" has 3 megabytes of free space after online defragmentation has terminated.

It doesn't make sense - theres no limitations but collectively the users mailbox's would amount to about 5GB - its weird?

Thanks for your response by the way

 

[ntinlin]

What about the dumpster deleted items column? What does that show.


Also what is the size of priv1.stm?

Neill

 

[ncolsgk]

If you did an offline defrag and it didnt finish, make sure you delete the temp file because it takes up a ton of space. At least thats what it did for me.

 

[58sniper]

Look for the other 1221 event. The one you quote is for the Public Folders.

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[xmsre]

For each store, and event ID 1221 is logged after online defrage completes. In the text will be the amount of whitespace. THe 1221 you site is for the public folders. Are there any more? MS recommends that online maintenance, including online defrag, complete at least once every two weeks. If this is not happening,then you need to enlarge the window in your is maintenance schedule.

Once you delete an item from a mialbox, a pointer is marked deleted for the deleted items retention period. At the end of that period, if no pointers to an object exist, the object reference is remove and the space the object once occupied is marked as free space. This happens DURING ONLINE MAINTENANCE. If online maintenance never completes, nothing will ever be marked deleted, no pages will be freed, and your database will continue to grow.

The process is:

1. Delete items
2. Wait for DIR to expire
3. Wait for online maintenance
4. Wait for online defrag
5. Verify freespace in the event 1221

Then, if and only if you have a large amount of free space that will neve conceivably be reused, you may consider an offline defrag if all of the below are true:

1. You are running standard or SBS and have no ability to create another store in your environment.

2. You are short on space

3. The whitespace will never fill again.


At a 135GB store, I'd bet you're running Enterprise. First, make sure the online maintenance, including online defrag completes. Next, create a new DB. After that move all the user mailboxes to the nesw db. Lastly, dismount and delete the old db.

 

[gmono]

Hi 58sniper-
The database "First Storage Group\Mailbox Store (W2K3JOR001)" has 118369 megabytes of free space after online defragmentation has terminated.

The priv1.stm is 210mb's and its SBS 2003 so I can't create another store.

 

[58sniper]

The database "First Storage Group\Mailbox Store (W2K3JOR001)" has 118369 megabytes of free space after online defragmentation has terminated.

Ouch.

Here is where you'll get different answers from different people.

Typically, I would say that if you're not hitting a hardware limitation, like drive space, to leave it alone. Exchange will just reuse the whitespace (actually, more effectively than if there wasn't whitespace). However, that's a LOT of whitespace, and the relatively low number of users aren't going to use that much whitespace any time soon (well, not likely, anyways). So, this is one of the VERY few times I would say it's probably a good idea to do an offline defrag. This will require some downtime.

Do a FULL backup. Verify it's successfull. And then, follow MarkDMac's FAQ faq955-5581 to do an offline defrag.

Keep in mind you'll need 110% of the database size as freespace before you start this. NO EXCEPTIONS.

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[gmono]

Does anyone know how to run the defrag when there isn't enough space to run it. I was trying to run it by creating the temp files on an external hard drive but I'm still getting the following error
Operation terminated with error -1808 (JET_errDiskFull, No space left on disk) a
fter 929.0 seconds.
This is the command I used - where e=external drive.
E:\>eseutil /d d:\exchsrvr\priv1.edb /t e:tempdfrg.edb /f e:\tempdfrg.stm

Any suggestions.

 

[zbnet]

I think I'd favour xmsre's suggestion in this circumstance.

 

[58sniper]

Well, for one, you're missing a backslash in the path for the .edb file.

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[xmsre]

If you can temporarially add another exchange server, even running in a virtual instance, do that and move the mailboxes. The new store will only be 19GB. After you're done, dismount and delete the original store. Create a new one and move the mailboxes back from the temp server. You can then back the temp server out of your site.

Although the process has more steps, It'll still be faster and require less space than than trying an offline defrage with /t and redirecting the temp file to a network share.

 

[58sniper]

I'm not so sure about that (and he's using an external drive, not a network share). I don't think standing up another Exchagne server (virtual or physical) is going to take less time and just doing the defrag. If you stand up another box, you're still moving 19GB of data to it (and back). An offline defrag is doing the same thing.

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[xmsre]

19 plus 19 or 38 vs 135 over and 19 back or 154. Esutil reads each page of the db, then puts them into tempdb. at the end of the process it deletes the original, renames the tempdb, and copies it back.

 

[xmsre]

For the other route. You install a temp exchange server. It's not holding the system folders, it's not the first server, and it's not a routing bridgehead and has no pf replicas. The install process takes maybe an hour. Movemailbox over and back, then just delete the server object from AD and reimage the box to cleanup. Maybe another 3 hours or 4 total. Compare that to 14 hours plus to run the offline defrag. With this process the store stays online and only a number of users equal to the number of simultanious movemailbox threads are impacted. With offline defrag, all users are impacted while the store is down.

 

[58sniper]

You're not copying 135 GB over using either method, since there isn't 135 GB of data.

Using a move mailbox method also loses dumpster data, and, if you're not carefull, Single Instance Store.

You still have downtime using either method, as you can move all 19 users at one time using Move Mailbox. That will certainly generate a fair amount of transaction logs. And, since you're moving the mailboxes twice (once to the swing server, and then back), you're generating the same data in TLs twice.

Pat Richard, MCSE MCSA:Messaging CNA
Microsoft Exchange MVP

 

[xmsre]

You do not lose SIS.

http://support.microsoft.com/kb/175481/en-us

You could loose the dumpster, but isn't that the point? I could set dir to 0 wait for dir , wit of online namaintenance, wait for oline defrage, read my 1221, - net effect AI lose the duimpdeter.

or I could jet move mailbox now - net effect lose dupster.


What;s the point other than a 2 day wait before losig the dimpster?

 

[Zelandakh]

Exmerge is your friend - exmerge the users out, dump the database, recreate and exmerge them back in. It will be a lot faster and a lot safer.

 

[xmsre]

except for rules and the other odds and ends that break.