Windows 2012 (with Dedupe) and large incrementals

[joesanto]

Hi Guys,


There could be a simple non-Veeam reason for this, but I'm struggling here...

We've had Windows 2012 with dedupe enabled for about 8 months now. We've never had a problem, but since last week, each incremental backup has been large and each replication seems to take a long time to consolidate its snapshots. I can't find any reason within Windows for the incrementals to be this large.

I'm still investigating the issue, but I was wondering if anyone had this problem? Perhaps someone has experienced this problem?


Thanks.

[foggy]

Joe, could you please clarify whether the amount of transferred data (and hence, the size of incremental files) has increased or just the time incremental job run takes?

[joesanto]

I'm not sure, because the incremental doesn't finish in time. We run in backup windows, because we can't run them during the day.

One thing I did notice is this:

(Note: the size of the VM is 3.8TB)

1. The average first incremental after a full backup will "read" 37GB and "transfer" 23GB.

2. Now, the first incremental after a full backup will "read" over 1.9TB!!! (could be more, but it failed) and "transfer" only 25GB.

[joesanto]

To clarify, the incremental takes much longer than it should.

[tsightler]

First of all I would manually verify if there are any "orphaned" snapshots on any of the VMs that are experiencing problems. It sounds like your system is no longer using CBT, any chance that you have added this system as a proxy to Veeam? Any recent power failures or other issues that have caused storage related issues? You may need to simply reset CBT on the VMs that are having this problem.

[foggy]

Yes, looks like CBT is not used on these job anymore, so you could try to reset it after checking what Tom has suggested and see whether this helps.