Replication error after changing source disk size

[mplep]

Hi,

I increased the size of a source VM's disk yesterday and the replication is now failing with "An attempt was made to access the data beyond the end of the virtual disk. Size of the disk". The target VM shows the increased disk size though. I asked on the forum whilst evaluating Veeam if increasing the disk size on the source would affect replication and the answer was no. We frequently resize and add disks so this is important to us. I'm replicating ESXi 4.0 to ESXi 4.1 using Virtual Appliance mode on Veeam 5.0 FYI.

I've searched and found some previous threads but these are locked. I've also noted the release notes extract below. But I had believed that making changes to source disks would not require re-sedding/re-setting up the jobs again.

Is this a support call then? I'm posting here first as I've already had and noted correpsondence on this issue.

Just one further question about the extract below. I have several VM's within the same replication job. Does this mean I have to re-seed ALL of them again? That would be a pain if so.

Replication job with ESXi target starts failing with “An attempt was made to access the data beyond the end of the virtual disk” error after source VM disk size is increased. To work around this issue, remove replica VM, and re-create the replication job after making changes to source VM disk size

Thanks

MPLEP

[Alexey D.]

Hello MPLEP,

You don't have to re-set/do a full run for your job. Just go to the screen where your replicas are listed.
Then expand your node, it will show all VMs. Choose 'delete from disk' for the VM with increased disk size.

During next run, the job will notice missing VM and will do a full pass for it, taking in account new disk size.
For all other VMs there will be an incremental pass. Hope this helps!

[mplep]

Hi Alexey,

Thanks for the reply. I had believed that altering the source disk size would not require a full re-seed of a target VM. For example, we have a file server which has lots of individual volumes and any one of those might require resizing as growth demands. Having to delete and redo the whole VM is a problem as our Veeam server is at of other end of a 10MB line and the source VM can grow to a few 100GB. The re-seed would take days during which we have no protection. I posted the question about resizing during my eval period http://www.veeam.com/forums/viewtopic.php?f=2&t=5353

Thanks

MPLEP

[Alexey D.]

Anton's answer to your post says 'changes will be applied' but say nothing about incremental runs.

This behaviour is by design, as stated in Release Notes. This was hard to implement - so we put this in RN.

[mplep]

Hi Alexey,

Antons reply says "Hi Mark, yes this should work fine" As my question was about an already replicated VM I thought it was fair to assume I was questioning behaviour in incremental runs. But regardless I need to try and fine the least painful way around this. Is it possible to just delete the individual VMDK on the replica that I am resizing and retain the others? Or is it just a complete start from scratch each time?

You say you have already looked at implementing this. So are you still considering this feature in the future?

Thanks

MPLEP

[Alexey D.]

All VMDK disks should be replaced, there is no workaround as for now. Maybe in future, but not in the upcoming update.

[Gostev]

We will consider this feature for the next release in which we are spending a lot of effort specifically around enhancing replication capabilities.

[mplep]

Thanks for the assistance. That would prove a very significant feature for us interally and also for our "service provider" service as well.

[as-ap]

Has there been any more work on this issue?
We just re sized a VM that is 550GB. I really do not want to have to re replicate that server over our WAN link if there is a way to just re size the replica.

[Vitaliy S.]

Currently, you should either re-seed this VM or perform a full replication run of this VM.

[JoeIanni]

[merged]

Hi,

I would like to know what the impact of increasing the size of the vmdk file, keeping in mind that this particular server is being replicated across to a DR site? Will the VMDK file at the DR site simply increase in size or do we have to reseed it again?

Any help would be appreciated.
Regards,
Joe