Barracuda Backup - Changed Block Tracking not working properly

[pjw73nh]

Greetings,

WHAT IS HAPPENING:

In an nutshell. I am trying to troubleshoot an issue with a Barracuda backup and Change Block Tracking.  It has been happening since the server (and the Barracuda 190) was installed a long time ago.

Long story short. With CBT enabled (Changed Block Tracking) Barracuda is only supposed to back up "changed disk blocks".  It does this at the virtualization level, not the guest level. In my case, it is backing up about 124 GB every night. (2-3 hours to the appliance), then 10-12 hours to transfer to the cloud.

MY SYSTEM:

There is only 1 VM (guest) on the host. The two drives on the VM (guest) are a 60 GB OS drive (SBS Server 2012 Essentials), and a 100 GB data drive.  Both have 33 GB used on each one (total 66GB).  Expected nightly changed data should be an absolute maximum of 6-8 gb each night much less ordinarily. 

Nothing on the host OR the  VM (guest) server adds up to the 124-128 gb that Barracuda says is being backed up.

HOW DO I KNOW IT IS NOT WORKING PROPERLY?:

I have two other systems in separate buildings that are working as designed. Thus, I have an A,B,C comparison.  A and B are working properly. C is not.   They are all the same hardware platform HP ML350p Gen8. They are all configured alike except for the sizes and quantities of disks in the RAID 5 array. They are all ESXi though the one in question is on 5.1.0  and the other two are on 5.5.0.

Not that it matters, but the guests on these systems are: Host System A has a brand new Server 2012  R2 STD .  Host System B has a "standalone converted" SBS 2011 .  Host System C has a brand new Server 2012 Essentials.

WHAT HAVE I TRIED?:

I have made certain all parameters in the settings have CBT enabled and set to true.  I have tried re-configuring a new backup selection within the Barracuda 190. I have been in touch with Barracuda 1st and 2nd tier support techs on multiple occasions. At first they told me it was a snapshot issue.  I looked at the snapshots as best I knew how (I am a noob, but I received some excellent info here as to how to look at snapshots, and running VM tasks from p_hall), then, when they couldn't tell me where in the snapshot phase the issue was,  They said "It's functioning as designed", (and that even though I only have 66 gb of disk space used, out of a total of 160 gb on the guest) "It's really not backing up 125 gb each night. it just reports that it is".  Yet my bandwidth usage says otherwise.   WHAT !!!!!!!!??????? 

I finally took the known good Barracuda from the FD (System B's Barracuda)  and put it at the PD (in place of System C's Barracuda) for a few nights.  It behaved just the same as System C's original Barracuda . It backed up approx 147 GB the first night and 128 the next two nights.

See attached for samples of the 3 systems backup reports from late last year. Yes, they're older, but the results are the same. 

Any suggestions as to how to solve this issue?  Thank you.

[pjw73nh]

Hello again,

In a little less than a week my post has been viewed 40+ times with no replies. This tells me either I have posted in the wrong place, given a poor description of the issue, or my issue is so unique than no one understands it, or no one has experienced it before.

Any suggestions?

Thanks.

[pjw73nh]

I would like to mention that my issue is now resolved. Still using CBT, my nightly backups are more in line with what is, was expected.  Instead of backing up 128 GB EVERY night (I still don't know where the figure comes from), I am backing anywhere from 3-15 GB per night.  Quite a difference.

The issue was solved with a simple  upgrade of ESXi from 5.1 to 5.5 upd 2.  No other changes made.

Would it be a fair and accurate assumption that contrary to VM KB 1020128 that CBT is, in fact, NOT supported on ESXi5.1?

I will be the first to admit that VMWare is not my forte.  That said, I am rather discouraged that there was not one reply to my post with any suggestions.  Did I post it in the wrong place. Poor choice of title/keywords? Was it poorly worded? Unintelligible?

I feel I was very thorough in my researching the issue.  I found this which clearly mentions CBT is supported in ESXi 4 and above.

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1020128

I was diligent in troubleshooting to the point where I ruled out just about everything except the Host OS.

Not knowing how big the community is, I would've thought that either there was someone out there either with my exact same issue, or that at the very least there was someone out there that was successfully running ESXi 5.1 with CBT on a Barracuda system. 

I am open to any suggestions that may make future posts/questions garner more replies.

Thanks.