Windows Server 2008 freeze during backup

[dcz01]

Hello,

i have a problem with an Windows Server 2008 Standard 32-Bit with an new TSM 7.1.3 client.
Sometimes the server freezes when using an txnbytelimit of 25600 direct when startet with the backup.
All other day it backs itself up without any problem or error.
Now i changed the txnbytelimit to 512000 in my clientoptionset on the server and then the server freezes completely every day... only an hard reset helps then to bring it up again.
The Server runs on an Fujitsu TX150 S3 hardware with old SCSI-HDDs, an Mylex RAID-Controller and 4 GB RAM.
McAfee VirusScan Enterprise its running on the server and the scanner is running every day at 21:00 to scan the whole system.
The system is used for an ERP system with the Windows Terminal Server feature activated.

Does anyone know a good solution?

 

[marclant]

Don't run the backup during the AV scan. Both are taxing the disks and competing for resources.

Is it hung or so busy that's it's unresponsive? Symptoms may look the same, but the former could be a code defect with Windows or a driver, the latter is due to a workload to high for the resources available.

How many files does the backup need to inspect?
If you watch CPU and memory when starting the backup, which of the two goes up too high first?
Lower the txnbytelimit to reduce amount od data kept in buffer before sending.

 

[dcz01]

The server is completly unresponsive and then after 60 minutes the session with the TSM server ist canceld by the server (idletimeout).
When this happens, i can't even log in direct on the server nor over terminal services. This happend in my test from now without the AV scan active.
So the AV scan can't be the one which kills the complete server.

The backup is inspecting about 90.000 files before starting to backup new files which where changed.
And when i start an manual backup on the server with the dsmc inc command, the server does it without error or getting unresponsive.
This is only happening in the scheduling mode with the services from Windows.

 

[marclant]

So the AV scan can't be the one which kills the complete server.

I wasn't suggesting that the AV alone killed it, just combined with TSM it might. But if you can recreate the problem without AV, than it's not the case.

The backup is inspecting about 90.000 files before starting to backup new files which where changed.

How many total to inspect, you'd see that in the stats at the end of a successful backup? If in the millions, may need to consider using memoryefficientbackup to reduce the amount of memory needed.

And when i start an manual backup on the server with the dsmc inc command, the server does it without error or getting unresponsive.
This is only happening in the scheduling mode with the services from Windows.

Unless the machine workload is different when you do manual versus a scheduled, this is odd because both are doing the same task, only difference is one is on the foreground and the other in the background. Also, when you do a manual, do you backup the same data as the schedule? do just "dsmc i" so that it backs up both systemstate and the filesystems, or just backing up the filesystems. The schedule is in most cases doing the equivalent of "dsmc i".

May want to try managing the scheduler with CAD, that way the scheduler gets restarted for every backup. Doesn't mean it will stop the problem from happening, but at least you know it starts from a clean slate every time.

 

[dcz01]

Well... now i solved the problem myself. It was only an overheating problem of the server why it shutdown every time...
Now with an new thermal grease an an completly clean case it works all fine.
Thanks marclant for the help and the good ideas.