HELP! Replica job changes disk format

[GabesVirtualWorld]

Hi
Replicating a Hyper-V (SCVMM) VM from a 2012R2 cluster to a 2016 cluster, also changes the harddisk type from fixed to dynamic. This is highly unwanted because when I need to change the disks back afterwards, it will take a lot of downtime.

How can I prevent this change in disktype? And why in heavens name, does it even happen?

Gabrie

[DGrinev]

Hi Gabrie,

This behavior is expected and cannot be changed in the current version of the product.
We have it on the roadmap to change the said behavior in the future versions. Thanks!

[foggy]

Hi Gabrie, could you please tell why do you want to keep them fixed?

[GabesVirtualWorld]

Hi,

The main reason we want to stay on fixed disks:
- We've discovered an UNMAP issue with our storage and dynamic disks on Hyper-V. Performance is heavily impacted when UNMAP is active.
- Our monitoring of the CSV volumes is currently based on fixed (and not growing) disks.
- For 1st line engineers it is quite a task to figure out if another VM can be added to a CSV when all VMs are thin.

Also a question for VEEAM: Why is the diskformat changed during a replication? Especially since it is nowhere to be found in the documentation that this change will happen and also the job creation wizard doesn't warn about this.

For the VMware Powershell cmd-let, there luckily is an option to force the disk type. But I also talked about this in the powershell section of the forum.

Can I use seeding to force the fixed format? Create first replica. Commit snapshot. Convert disk. Create new replication job and use the target as a seed?

Gabrie

[foggy]

Seeding might serve as a workaround, worth a try.

[GabesVirtualWorld]

Nope, after making the first sync and committing the snapshot. I used that VEEAM as a seed and started a new replication. After the replication finished and removing the snapshot from the replica, I now have two VMs and both on dynamic disks. Quite a different behavior from VMware replication.

[foggy]

Do you mean the second VM was created when you mapped another job to replica with switched disk?

[GabesVirtualWorld]

I reran the test and I think I messed up in the first test. The second time it worked as expected.

Original- VM (fixed disks) --> Replication to ReplVM
ReplVM is dynamic with differencing disk (Snapshot).
Remove Replica Job.
Commit snapshot on ReplVM.
Convert ReplVM to fixed disk.

Then create new replication job to create replica from Original-VM to new VM and use ReplVM (with fixed disk) as replica (seed).

End result:
Original-VM = fixed
ReplVM = fixed with snapshots. Strangely enough the ReplVM was not renamed to the new suffix in the job, but kept the name of the mapped VM.
I then performed the final failover and I now have a running replicated VM with fixed disk.

Sadly though I do have some extra work with converting and then re-seeding, but for now it is a workable alternative.

[migmol55]

Hi

When I configure a VM in Hyper-v for replication VEEAM VBR v10 change the disk format for Dynamic Disks.

Is it possible VEEAM in replication doing the Failover / Planed Failver action maintain disk format as FIXED Disks?

Is there any advanced configuration in VEEAM Backup and Replication v10 to maintain disks format?

Thank You.
Regards.

[]s MM

[Egor Yakovlev]

Hi Miguel,
I have merged your post to the existing thread.
This is expected behavior and workaround above with seeding should be valid still.
/Cheers!

[mausger]

Is this still the same behaviour in Version 12.1.1.56, because we have the sam issue.
You wrote, it is on your roadmap in a post of 2019 so after 5 years is it still not implemented to set the behaviour.
We also want the same disk type on the target so we do not have additional work, and by the way you need to have extra disk space when converting because the new VHD file is created first as new one and after that you can delete the old one.

Regards

[Egor Yakovlev]

Hi Martin.

No changes here.
Is there a particular reason to perform conversion back to fixed besides consistency with the source machine?

p.s. as far as I am aware, converting disks to fixed on ReFS will be using block cloning.