Exchange Logs... possibly causing perf issues

[exchangeboy]

Hello,

I've been battling performance on my one of my Exchange servers for about 2 weeks now. My backup has been running off and on lately causing log files to be kept for almost 4 days at some points. When these log files are stuck and we are having Exchange performance errors, the problem is usually lifted by waiting for the backup to run properly that night. If the backup still has issues before a day ends, we have been moving them off the drive, and deleting them manually at a later time.

My question is, if the Exchange services are not stopped and the log files are moved/deleted, will this cause further issues down the line when Exchange tried to do an online fragmentation? Basically, I've found out that the people here have been moving/deleting stuck log files that should have been purged with the nightly backup and NOT STOPPING the Exchange Services.

Lately, we've been having issues of mailbox stores dismounting themselves and many ESE performance errors. Here are some examples:

Event Type: Error
Event Source: MSExchangeIS
Event Category: General
Event ID: 9558
Date: 4/9/2007
Time: 11:23:37 AM
User: N/A
Computer: xxxxxxxx
Description:
An error occurred while writing to the database log file of storage group "First Storage Group". Attempting to unmount all databases in this storage group.

For more information, click

http://www.microsoft.com/contentredirect.asp.

Event Type: Warning
Event Source: ESE
Event Category: Performance
Event ID: 509
Date: 4/17/2007
Time: 3:37:43 PM
User: N/A
Computer: xxxxxxxxx
Description:
Information Store (4528) Second Storage Group: A request to read from the file "F:\Exchsrvr\mdbdata\Mailbox Store 9.edb" at offset 29481672704 (0x00000006dd3ea000) for 4096 (0x00001000) bytes succeeded, but took an abnormally long time (68 seconds) to be serviced by the OS. In addition, 4 other I/O requests to this file have also taken an abnormally long time to be serviced since the last message regarding this problem was posted 8739 seconds ago. This problem is likely due to faulty hardware. Please contact your hardware vendor for further assistance diagnosing the problem.

Event Type: Information
Event Source: ESE
Event Category: Online Defragmentation
Event ID: 704
Date: 4/19/2007
Time: 6:00:00 AM
User: N/A
Computer: xxxxxxx
Description:
Information Store (4528) First Storage Group: Online defragmentation of database 'E:\Exchsrvr\mdbdata\Mailbox Store 3.edb' was interrupted and terminated. The next time online defragmentation is started on this database, it will resume from the point of interruption.

Event Type: Warning
Event Source: MSExchangeIS Mailbox Store
Event Category: Background Cleanup
Event ID: 9791
Date: 4/19/2007
Time: 3:13:25 PM
User: N/A
Computer: xxxxxxxx
Description:
Cleanup of the DeliveredTo table for database 'Second Storage Group\Mailbox Store 9' was pre-empted because the database engine's version store was growing too large. 0 entries were purged.

For more information, click

http://www.microsoft.com/contentredirect.asp.

If you have any input on this, I greatly appreciate it!

Anthony

 

[SkreeM]

Deleting the log files is always a bad idea! In the case of an issue with the databse uncomitted logs can be replayed to recover the messages etc, if there are logs missing etc you will have major difficulties recovering.

I would invesigate the issues with your backup as primary to the problem ... If you cant resolve it use NTBackup to backup the information store to a file on disk or over network every day, that will ensure the logs are flushed then backup that file to tape/removable hd etc.

Once you have a relliable backup running etc you can look into taking the databes off line and doing checking / derfagmentation etc.

It appears you may have issues with the I/O on your disks have you checked them?

Skr

 

[exchangeboy]

The backup issues are in the works for being repaired, it has been about 3 days and we haven't had any hiccups. It was possibly forgotten about and no tapes were fed into it, and it caused the backup to stop during the wait for a new tape. We've now appointed someone to constantly check the backup.

I haven't been able to check the I/O for the disks because I don't know how to. How can I check the I/O for these disks?

Thanks!!

Anthony

 

[DJCopa]

Performance Monitor on Exchange server should do the trick - Just add the respective Snap-In and you can add Disk Reads,etc to your hearts content.