Changed Block Tracking & moving to a new datastore

[withanh]

Based on what I'm seeing, I pretty much know the answer to this, but I would like a confirmation (or a rebuttal with a more correct answer).

We backup using CBT over SAN mode. In anticipation of migrating to Exchange 2007, I moved my Exchange 2003 VM to a different datastore. My Exchange backukps usually run about 2.5 to 3 hours, but right now I'm at almost 10 hours and it says there are over 4 hours remaining. I'm guessing that this behavior is because *all* blocks were changed when it moved to the new datastore and as a result VBR is having to re-read all the blocks to see if they really changed or not.

Am I understanding correctly what is going on?

[Gostev]

Probably not, changes are tracked on VMDK level, not on storage level. So either datastore is really slow, or the job has failed over to network, or something else.

[withanh]

Not the datastore, it's a Raid10 on 4 x 300GiB x 15k RPM SAS drives and there's nothing else on that spindle set.

It could be network mode...when I look at the drive configuration on my Veeam server, I can see that one of the drive arrays shows as basic/healthy/primary but also shows as Read Only where the others show as Online. In theory that shouldn't matter since I'm only reading data anyway, but that is one difference that I can see. Is there a way to go back and see which spindle set this drive is associated with?

[withanh]

Screenshot at http://www.flickr.com/photos/45779723@N04/4203894252/

[Gostev]

Not sure... I know very little about storage

[withanh]

Well, that was weird for sure, it did the same thing on last night's backup. All the iSCSI connections looked correct, but I decided to reboot the server and tried the job again and it is running in SAN mode now not network mode.

Oh well, at least it's back up to speed!

[Gostev]

Great Thanks for the update!