Unable to write to a shared folder while it is backing it up

[Jay_Ryan]

Here is a little information before the problem. My DBA has local backups running on each of his SQL servers that dump into a shared folder on our backup server's array. This includes a full DB backup each night and transaction log backups about every 20 minutes throughout the day. The folder has multiple subfolders as well. This is our second backup server which is running 11d. We have another backup server running 9.1 which we use to do this on and it worked fine. When we moved the share folder over to the new server all of sudden we have problems. When Backup Exec is backing up the shared folder it does not allow anything written to that folder which fails the SQL backups trying to go through at that time. We sometimes get a "Delayed Write Failed". I have all the same settings on 11d as I did with 9.1. I'm puzzled with this and I've been banging my head on my desk. Any ideas?

[perry_baker]

I'm not certain the problem is relative to Backup Exec.

If you moved the shared folder to a new server then the different device could be the issue. It could be due to the scsi card (or other controller card) to which the hard drive where the backup-to-disk folder is located.

But BE 11 has a new feature to provide the ability to restore AD, Exchange and Sharepoint data via full DB backups, this is known as GRT (acronym for Granular Restore Technology), so there is a possibility that you are encountering something relative to the new feature.

You could test this possibility by not running your BE backup and instead attempt to copy a similar amount of data to the shared folder while the SQL dump is supposed to also be accessing the shared folder. The idea is to mimic the process of moving data from the BE media server to the shared folder while the SQL dump is taking place. If the same problem appears then you would want to examine the subsystem to which the hard drives are attached.

If the problem does not occur and your running a backup of Exchange, Sharepoint, System State / Shadow Copy Components, try removing those resources from the backup and see what you encounter. The idea is to get the backup to a point where it is just simple file data so that the new GRT eligible data is not included in the backup and determine if the problem exists.

[Jay_Ryan]

Sorry for not replying to this sooner. Yes, we figured out what it was. A drive had gone bad on our array and degraded the RAID. I guess it had to do something with heavy I/O.