VMFS3 to VMFS5 and Veeam

jenuster

Limp Gawd
Joined
Mar 8, 2011
Messages
154
I have a Exchange VM in particualr that lives on a local datastore and by default was formatted so that the most space I can allocate to each hard disk is 256 GB (it's running ESXi 4)

I'm planning to deploy a new host and install ESXi 5 which has lots of storage and use Veeam to replicate the Exchange VM to the new host. Now the question is when I replicate that VM from a VMFS3 volume to the new server which has VMFS5 volume, will the block size change to 1MB?

The idea is to have the ability to expand one of the drives on the Exchange VM past the 256 limit once on the new ESXi host.

Thanks.
 
the VMFS 5 Volume's Block size will be 1MB.

are you planning on running both at the same time when you use the word replicate iim thinking more of a DR solution for this instance, or are you just backup/restore run on VMFS5?

Problem is if you are doing bi-directional Replication this is not going to work well if at all. if your end goal is to get it migrated i would just say get it migrated over to vmfs5 and not replicate..

more information would be needed to really refine the answer..

block sizes for the VMFS Volumes are dependent on the version all VMFS5 volumes are block size 1MB (but it can allocate up to the 2TB marker)

VMFS 3 on 1MB block only goes 256GB
 
Sorry I wasn't clear enough.

The reason why I'm moving the Exchange server is that the disk on the VM is almost full. Unfortunately I can't expand it anymore (it's at 250 GB now) or start moving mailboxes because there is barely any storage left on the server.

The goal is to get the Exchange VM from the host running ESXi 4 to ESXi5 in the hopes that it will be formatted with the 1 MB block and will give the ability to increase my drive size more than 256 GB.

Everything is local storage, so i figured the best way to get the VM from ESXi 4 to ESXi 5 was to use Veeam as opposed to shutting down the VM and migrating it to the new host over the network.
 
My personal Thought for this would be to down the Exchange box and clone it over to make sure all goes well. it will also help with the migration speed, but i know in a real world this is sometimes not possible.

Maybe Iop might have idea on this one... ill message him to see what his thoughts are
 
Clone is my personal opinion on this one. Shut it down and clone that sucker. Will be a lot easier :)
 
Thanks, how sure are you that when I clone the VM to the new ESXi host, that it will be formatted with the 1MB block allowing my virtual hard disk to reach up to 2 GB?
 
Because the block size has nothing to do with the file. It's the underlying addressing scheme of the filesystem - when you put something on VMFS5, it only has one way to address it ;) Not gonna go into the structure to explain the limitations in prior versions, but it actually didn't have anything to do with the file as much as how many bits you had to address with.
 
shut it down and use the migrate feature of esxi it will move it and you will not have a problem. Assuming you don't have enterprise esxi licence.
 
Just have the poor persons licensing for ESXi here. I went from ESXi 4 to ESXi5, with intention of taking advantage of the VMFS enhancements, and use Veeam's "Quick Migration" option.

Worked out fine. Basically throws in a snap shoot, replicates the mess, powers down source VM, replicates over the snap shot (changes from original replication) then spins up the VM on the new host.

Not much down time, power down/small data transfer/power up.

Regarding the drive size limits, over my head there.
 
I usually just storage vMotion, but as Marty said, you can use Veeam. Once it's over to the new storage, you can always create another drive and migrate the DB. Of course you'll need to take it down to do that, but it really shouldn't take that long, rather then expanding the drive. Also, even with vmfs5, you can't go over a 2TB - 512 vmdk and if you ever intend on snaping it for backups or replicas, keep it around 1.9TB.
 
Back
Top