Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

URGENT HELP - IS shutting down at 16GB limit

Status
Not open for further replies.

Norwich

MIS
Mar 3, 2002
336
GB
The Info Store is shutting down due to hitting the priv.edb 16GB limit. We ran ESEUTIL to defrag the store two days ago so there is very little free space available (about 50 MB using eseutil /ms). The problem seems to be that the deleted mail retention time was 2 months (!) But as the system keeps shutting itself down when started, we can't get the auto-maintence to clean out the deleted mail with a retention period set to 0 days.

Any ideas on forcing the deletion of the retained deleted mail VERY VERY welcome.
 
Did you install exchange from a microsoft select cd? If so, run the SrvRmax.exe program file from the Select CD to install the Enterprise Edition of Exchange Server 5.5.
 
Hopefully you've got a recent backup.

If so, pull the server off the network, start the DS and IS but leave the connectors stopped. Bin your own mailbox. That should bring the size down a bit.

Clean each mailbox to minimise size. Then change deleted item retention to 5 days or something and defrag the database.

Once all of that is done, restore your mailbox. You'll probably find several GB...
 
Thanks Zelandakh, but I can't start the IS at all. The system detects the size and utomatically shuts it down.
 
Do you by chance have Circular logging enabled or disabled? If it is enabled you probably have several log files out there. Can you possibly delete any log files? Is there any way you can free up any disk space to allow you to start the IS and DS?
 
We do have circular logging, which files would be suitable to delete?

Thanks for this help
 
Wait, Is this a disk space issue or a PRIV.EDB size issue? If it's the size of the PRIV, then as Zel says, make sure you have a backup, stop all your services and run an ESEUTIL in Defragmentation mode. This should at least net you a few hundred MB's of database space which will allow you to start all the services again, start deleting stuff, and then run another defrag....
 
Ok,

I'm assuming you do not have an exchange aware backup program? If you do have an exchange aware backup program but you are running DIFFERENTIAL backups then the log files do not get purged. They only get purged when you run a Full Backup or Incremental.
(1) If you do have an exchange aware backup program you may want to run a full backup before completing this step.
(2) Check to see if there are several log files in the following directory
==>exchsrvr\mdbdata
* DO NOT DELETE EDB.LOG
* If there are tons of log files that start with EDB for example: ==>edb01bae.log These log files start at 5120KB and if you have a ton of them your disk space can get used up quickly. I would delete these but BE CAREFUL not to delete anything else. Your PRIV.EDB and your PUB.EDB reside here.

Let me know how you make out. Please use these steps with care.
 
Hi Bronto,

It appears he ran the eseutil /d program already. Hopefully I'm not giving bad information. But i'm guessing the priv.edb is not at its max but if there are several log files could this not increase the IS size? I did mention he should definitely backup but if he does not have an exchange aware program then the logfiles are probably going crazy and is causing a disk space issue.

 
Hi,

Their aren't that many logs that could be deleted for much gain.

For brontosaurus, it's the size of priv.edb (16,777 MB) that's causing the problem (it's shutting itself down sue to the size). I eseutiled 3 days ago but the deleted retention only meant abot 5 GB was recovered. Now eseutil /ms says that there's about 60 MB to recover !?! but I'm doing that now anyway to see what it truly brings it down to.

 
Hi,

While the system is backing up / processsing etc, I'll ask what is the best way to force the system to perform the nightly houskeeping at any time - specifically the hard deletion of the mail expired past the retention period?

I'm assuming to set the system to all maintenence at any time of day and just wait would do but can it be fired off manually?

thanks...
 
Egads, I'm sorry Zoey, maybe I'm confused (maybe?...).
Norwich, tell us how you make out after the process finishes. I'll be shutting up now.
 
Hey Bronto,

You are right as usual. I did not realize he was talking about the Priv.edb and took a shot at maybe the log files were piling up. That is some humoungous IS you have there Norwich. You may want to consider putting storage limits on your mailboxes if that is politically possible at your Company. As for the hard deleting question I assume you must have Deleted Item Retention days setup in your Private Information Store properties. You could uncheck the "Dont permanently delete items until the store has been backed up" so when your users delete their deleted items bucket it will be deleted. So there are basically 2 ways to do this.

(1) Place Storage limits on your mailboxes to manage the size of your IS better. You would have to decide mathematically how much limits for each mailbox.
(2) If you do step 1 you may not have to remove the Deleted Item Retention. You may want to keep this in case someone deletes mail and then they come to you asking to recover. But of course that is your call.

Bronto,

Please don't ever shutup. None of us would learn anything. Every opinion you offer is usually right including this one. Please let me know if the 2 options I gave Norwich are good options for him.

Thanks
 
Hi,

Well, we implemetned the limits recently which most accounts are now under - the problem was that the store was soooo much larger than the sum of the component mailboxes. we had hoped the usuutil defrad we did the other day would take care of it but not by much. It's now reached the 16GB fixed limit and I now think it was because the previous Sys Admin had the retention for deleted messages set to 60 days - so all the frantic deletions by users with the new prohibitive limits did not have any effect.

My worry now is that as we did the defrag so recently, there will be very little to reclaim - hence the urgency in getting the system to clear out the old messages once the retention period is set to zero.


Thanks
 
Thanks Zoey, that was very kind of you. I think Mr. Norwich should indeed consider storage limits, and should maintain some deleted item retention, after all, it will save your backside in many instances. And, if I may be so bold as to offer one more suggestion? ENTERPRISE EDITION!!! :) I know, it costs money, but sometimes you have to spend a little to gain a little...
 
Enterprise will solve the problems assuming you have the HD space.

If you can't get hold of it, run the defrag anyway as you may only need a few MB under to start the ball rolling...its your only option at the moment.

Then you'll need to get some working space so clean all the mailboxes and bin your own mailbox (check your backups first)....
 
Bingo - got the thing working.

Defragged the database but as it had been done recently, it only freed up 40MB. However, this was enough to stop the IS service chocking on startup. Set the retention of deleted items down and turned on maintenance for the current time. At the quarter hour, it started deleting the non-retained items in the private store (event 1206) and finished after two hours. Degrag then shruck it down to a more manageable 6GB.

Thanks everyone for the help - v.much appreciated.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top