For the last month or two, I've been experiencing random crashes on our Exchange server while backing it up. It doesn't happen every night, and it doesn't necessarily happen at the same part of the backup job, but as yet I have no real answer.
The server in question is an HP DL380 RSO unit running with an MSA500 array. The primary server has 4GB of RAM and the backup has 1GB. There are no local disks in these two machines as such, all the disks are in the MSA unit. The server is running Windows 2003 Enterprise, with Exchange 2003 Standard. It is being backed up with BackupExec 12, with all the patches released to date, and on several occasions I have removed and reinstalled the remote agent to make sure it is using the latest verison.
My main issue is that what appears to be happening is that the primary server gets so overwhelmed when the backup goes wrong that the backup server cannot see that it is still running, and then takes control of the disk array. This means that the primary no longer has access to the disks and cannot write any error messages about what is going on! Eventually the server resets and I end up with the primary rebooting to the recovery screen asking me if I want it to reassume control of the array. For a while, the secondary unit seemed quite happy to run the backup, but now it has started doing the same. I did think that maybe there was an issue with the array controller in the primary server, but now I'm not so sure.
On the HP forums, people with similar issues have suggested making sure that all the drivers and firmware are the latest and greatest versions, which I have done. As mentioned above, I have reinstalled the remote agent as it has been suggested that different versions of the agent on the backup and remote machines can cause grief. On one occasion, this seemed to cure it for a week, but then the trouble resurfaced. I have since seem on this forum that it is suggested running AOFO and backing up Exchange is not a good idea, so I have made sure this is not running at the same time, but does anyone else have any suggestions at what else I can look at if this doesn't work? I'm fed up with having to get up early every morning to check to see if we have a working mail server!!
