Thursday 12 November 2015

Exchange Stall: Purging Exchange Logs

We’ve got our SBS (Small Business Solution) set up for client’s all-in-the-Cloud experience.

What we’re finding is that some vendor’s systems don’t trigger VSS within a VM running Exchange thus leaving Exchange thinking it is not being backed up. This means that the Exchange VM eventually stalls due to no space on the partition hosting the logs.

To remedy that run the following in a script say once a week, or more depending on the volume of mail, on the Exchange VM:

  1. Elevated CMD
  2. DiskShadow
  3. Add Volume C:
  4. Begin Backup
  5. Create
  6. End Backup 

Once the snapshot completes Exchange will think it’s been backed up and consolidate all of the logs.

NOTE: Make sure the VM/Server is backed up!

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book

2 comments:

Unknown said...

While I appreciate that this is a workaround, I'd like to see an article from Microsoft with an actual fix. I've lost track of the number of virtual Exchange boxes that don't commit logs after a backup from the Hyper-V Host. Seems to happen regardless of using Native Windows Backup or 3rd Party Utilities (Veeam, Backup Exec etc) I've actually lost track of the many things I've tried to resolve in the past and how many times I've looked at "vssadmin list writers" results just to see there's no reported issues there and nothing useful in the event logs.

VSS has been a mystical process for far too long and it decides to hiccup over the most ridiculous things.

This particular issue with log commits and bizarre backup problems was just one more reason I pushed for moving a lot of on-prem Exch boxes to Office 365 (where appropriate of course). The ugly cynic in me wonders whether that was the plan all along.

Philip Elder Cluster MVP said...

I've found that the problem is not in Exchange but in the way the backup software initiates and runs the backup.

We don't use built-in Windows Server Backup for any backups anymore. We use ShadowProtect.

If we set up a FULL/INCREMENTAL backup we don't have issues. But, if we set up a CONTINUOUS INCREMENTAL backup with weekly VSS snapshots we find our Exchange VMs are not necessarily triggered to consolidate.

What the difference is I'm not sure.

For now, we run this script. It works, and keeps things under control.

FYI: We don't do O365 anything.