3

Every night at 02:59 the following events are logged by my Exchange server:

Event Type:   Warning Event
Source:       MSExchangeIS Mailbox Store
Event Category:  Background Cleanup 
Event ID:       9871 
Date:           5/20/2009
Time:           2:59:59 AM 
User:           N/A
Computer:       EXCHFC2 
Description: There is an online maintenance overlap for database "First Storage Group\Mailbox Database". Please revise the online maintenance schedule.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.

.

Event Type: Information
Event Source:   ESE
Event Category: Online Defragmentation 
Event ID:   704
Date:       5/20/2009
Time:       2:59:59 AM
User:       N/A
Computer:   EXCHFC2
Description:
MSExchangeIS (12420) First Storage Group: Online defragmentation of database 'C:\Program Files\Microsoft\Exchange Server\Mailbox\First Storage Group\Mailbox Database.edb' was interrupted and terminated. The next time online defragmentation is started on this database, it will resume from the point of interruption. 

For more information, click http://www.microsoft.com/contentredirect.asp.

This is Exchange 2007 running on Windows 2003 x64.

Here's my Exchange maintenance schedule:

First Storage Group\Mailbox Database: 20:00 - 03:00  
Second Storage Group\Public Folder Database: 03:00 - 06:00

Note that I tried spacing them out more (First ended at 2, second started at 4) to no avail. The error message always appears a second before completion of the First Storage Group's maintenance plan.

The interrupted defragmentation noted in that second event was begun at 02:15. Another defragmentation of the exact same database was begun at 23:04 and finished at 02:08. Immediately before this full defragmentation, the interrupted one from last night was completed.

So it's as if the system is defragging the database in a loop, then complaining when that defrag is interrupted by the conclusion of the maintenance window.

How can I fix this behavior?

EDIT

My backups run at 06:30. There are no scheduled jobs on this machine.

The Best Practices Analyzer reports nothing relevant.

I understand that this is not a critical error, but I'm concerned that if I simply ignore it I'll overlook a future problem when maintenance is NOT being adequately performed. In that vein, I'd like to resolve whatever is causing the defrags to loop during the maintenance window.

2
  • Lets bounty it.
    – sh-beta
    Nov 9, 2009 at 19:10
  • Only two hours left! Is this beyond all hope?
    – sh-beta
    Nov 16, 2009 at 16:48

3 Answers 3

1
+50

The 9871 error is thrown when a databases window for maintenance is closed and it is still running transactions. It will complete the maintenance and its more than likely not actually anything to do with running the maintenance on the other database at all, its a pretty badly written error message.

Often this message is triggered when ba backup is running when the maintenance window ends, could this be happening?

It may be worthwhile to run the Exchange Best Practices Analyser against this server just to check if there are any problems it can pick up.

2
  • Answers to these questions added to my original post.
    – sh-beta
    May 21, 2009 at 14:37
  • 1
    This answer was auto-selected when the bounty expired. There is some useful information here that will hopefully help others but, unfortunately, it doesn't answer my question.
    – sh-beta
    Nov 17, 2009 at 22:16
0

Sam is right... it is just a poorly written error message.

Event 9871 is logged when the maintenance windows is over for a database, but a transaction is still being processed. Exchange doesn't end the transaction just because it hits the end of the maintenance window, it just logs the 9871.

You can safely ignore the error (or extend the maintenance window if you want to avoid the event being logged).

1
  • Extending the maintenance window does not avoid the event-logging for reasons outlined in my post (ie, it continually restarts).
    – sh-beta
    May 21, 2009 at 21:36
0

So it's as if the system is defragging the database in a loop, then complaining when that defrag is interrupted by the conclusion of the maintenance window.

How can I fix this behavior?

This is working as designed. From: http://technet.microsoft.com/en-us/library/aa996226(EXCHG.65).aspx

It is possible that this process does not complete before the time indicated by the Microsoft Exchange Information Store service. In such a case, the ESE engine stores the last table successfully processed. At the next occurrence of Online Defragmentation, ESE selects the next table in the metadata, and will resume processing tables until the maintenance window ends.. If online defragmentation can process all tables before the maintenance window ends, it will restart at the first table listed in the metadata and loop on all tables until the maintenance window ends. The changes made to the database are transacted. Therefore, a large number of delete operations, followed by an online defragmentation cycle, will generate a large number of transaction logs, because records are compressed into the fewest number pages.

Which means that it will loop through the databases during the maintenance windows.

As for the Maintenance Overlap message, the suggested is posted here: http://technet.microsoft.com/en-us/library/aa995795.aspx

By default, the mailbox store maintenance processes run nightly from 1:00 A.M. to 5:00 A.M. Because of the server resources required for the maintenance processes, it is recommended that you modify the maintenance schedules so that the mailbox stores have schedules that differ by at least 15 minutes.

So according to that if you start the Second Storage Group maintenance window at 3:15 instead of 3:00 the message will go away.

1
  • From my original post: "Note that I tried spacing [the maintenance schedules] out more (First ended at 2, second started at 4) to no avail. The error message always appears a second before completion of the First Storage Group's maintenance plan." I appreciate the confirmation that this is working as expected, but my goal is to fix the error message.
    – sh-beta
    Nov 9, 2009 at 19:38

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .