A few questions...

[jgeuder]

Hello,

I am currently evaluating Veeam right now… and I have a few questions I am hoping to get answered.

1)If I use CBT with a VM that gets backed up once per day and replicate the same VM using CBT once per week… wont the CBT information be useless to the replication job since so much time has passed and the changed block information is outdated?

Doesn’t the CBT get “reset” after each “incremental” backup so the next pass looks at only the changed blocks?

It seems like the backup and replication jobs would clash with each other.

If I am wrong, can you please explain exactly what is happening to get both of these jobs to use CBT?

To me this seems like backing up one Exchange server with two different backup products… logs would be missing – it would be a real mess.

2) Will CBT work if a VM is powered off?

3)In a previous post I asked about VSS – I have been a victim of corrupt AD and SQL using the old VMware SYNC driver – but since using the VMware VSS in ESX 3.5 U2 and VCB– I have had ZERO problems – why is Veeam VSS so much better – can you provide technical details?


I probably will find more questions….

Thank-You very much for your answers!

-Jason

[Gostev]

Hello,

1. No issues with CBT in this, or similar scenarios.
2. Yes, VM power state does not matter.
3. See here for example: VSS, just a random thread - there are multiple other existing threads with bigger discussions of differences between VMware VSS and Veeam VSS.

Thanks!
Anton

[jgeuder]

Hello,

1. No issues with CBT in this, or similar scenarios.
Can you tell me why\how?

2. Yes, VM power state does not matter.
Are you sure?
I tried backing up the same powered off VM 3 times in a row with one job - each time it took 3-4 hours. But works fine with a VM that is running.
I am NOT a developmer but found this when searching the internet for answers:
(http://www.vmware.com/support/developer ... notes.html)
Under Particular Conditions queryChangedDiskAreas and getChangeID fail
The queryChangedDiskAreas and getChangeID methods rely on support from the vSphere platform to examine file data in a read-only mode. The underlying platform cannot gather the requested data if the following conditions are true:

The query is regarding a virtual machine that is powered off
Information about the versions of the virtual machine being compared are on different ESX hosts
Information about the snapshot being compared is accessed using SCSI hot add transport mode


Copyright © 2009 VMware, Inc. All rights reserved.


Sorry to be a pain - just deciding on Veeam by end of year. just looking for a little bit deeper technical dive into the product.

Thanks!

[Gostev]

1. Every backup or replication job maintains it's own changeID, so there is literally no way that they can affect each other. For every incremental pass, every job queries changed blocks data with changeID preserved on the last pass, and gets the list of all changed blocks since the provided changeID sequence number.

Basically, VMware CTK is no different from change tracking systems that existed out there for years - like Microsoft Active Directory, Exchange server and other. Exactly the same concept. None of these systems can be affected by multiple clients requesting changes data since various points of time obviously.

2. I will ask devs/QC to confirm.

[jgeuder]

Anton,

Any feedback from devs\QC on my "Powred Off" CBT question....

Thanks!

-Jason

[jgeuder]

???

[Gostev]

I did not get a response to my email from devs. I will ask again when I am back from vacation. Sorry about that.

[Gostev]

Jason, we were unable to confirm this issue in our lab. We found no differences with CBT as it comes to Powered Off VMs. If you'd like, please open support case so that we could investigate your logs to determine why you are having this issue. Thank you!