EX2007 GRT backup fails

[JRV]

Migrating from EX2003 SP2 to EX2007 SP2. EX2007 Storage Groups and Databases are set up but mostly empty at this point. LCR is in use, and backups are made from the passive copy if available, else the active copy. GRT is enabled. All SGs are backed up by one Selection Set.

Physical (& logical) disks (addressed as mount points) are as follows:

Disk1 SG1/DB1
Disk2 SG2/DB2
Disk3 SG3/DB3 + SG5/DB5
Disk4 SG4/DB4 + SG6/DB6

Disks 3 & 4 have each SG in a separate folder, but they share the same spindle. The 4 passive LCR disks are laid out the same way.

SG1 & SG2 back up without problems. The remainder all throw this error:

V-79-57344-891 - Using the job settings, Backup Exec attempted to back up the passive copy of the Exchange database. However, the passive copy was not available after the Microsoft Volume Shadow Copy Service (VSS) snapshot was performed. Try to run the job again.

Writer Name: Microsoft Exchange Writer, Writer ID: {76FE1AC4-15F7-4BCD-987E-8E1ACB462FB7}, Last error: The VSS Writer failed, but the operation can be retried (0x800423f3), State: Failed during backup complete operation (12).

V-79-57344-65233 - AOFO: Initialization failure on: "\\ServerName.DomainName.local\Microsoft Information Store\StorageGroup.3". Advanced Open File Option used: Microsoft Volume Shadow Copy Service (VSS).
Snapshot provider error (0xE000FED1): A failure occurred querying the Writer status.

Check the Windows Event Viewer for details.

Writer Name: Exchange Server, Writer ID: {76FE1AC4-15F7-4BCD-987E-8E1ACB462FB7}, Last error: The VSS Writer failed, but the operation can be retried (0x800423f3), State: Stable (1).

The Symantec KB article suggests it's an intermittant prroblem, but this happens every time. Not sure it's talking about the same issue.

Only commonality I've come up with is that there are 2 Storage Groups and/or Databases on the same physical or logical disk where the backups fail. And if I back up SG3 & SG4 in one job, and SG5 & SG6 in another, there's no problem, which tends to support that conclusion. However, that was not an issue on the WS2003/EX2003 server, but maybe it is with WS2008/EX2007.

I know it's not a best practice to have 2 SGs share a physical or logical disk, but this is a pretty small server, and it is a supported config. Can anyone confirm that the problem is that 2 SGs share a disk? If each SG had its own partition, would the problem go away?

[Colin_Weaver]

 Suggest you switch the job to Backup from Active Copy ONLY

As you are running LCR you are not gaining that much from trying to backup from the Passive Copy - i.e. same CPU, memory pools etc. And as Symantec have acknowledged an issue that they are liaising with Microsoft with regarding passive copies - to get regular reliable backups use the Active Copy.

Doc for Passive Copy issue is http://entsupport.symantec.com/docs/305815
And I know your error code is slightly different but that might purely be because you are running Exchange 2007 SP2 - which is not fully certified by Symantec yet, so therefore may not have been checked for the Passive Copy issue.  

BTW I have a feeling the intermittent status is because some environments see 1 backup in 10 fail, that does not mean some other environments do not see 9 backups in 10 fail. 

[JRV]

Probably a reasonable workaround; I had considered that. Just wanted to learn a little more about the issue, first.

Backing up from the passive copy does isolate the disk activity to the passive spindles. But as I say, this is a small server, and if their ancient EX2003 box could handle backing up from the active (and only) copy, surely we can handle it on the new hardware!