Busted VHDX - Data Drive on File Server (hyper-v svr 2016 VM) (original) (raw)

February 9, 2025, 9:58pm 1

Hey guys, me again…
Am stuck (although i believe i have fall-back plans)

For ref: no User activity since Friday, and none will login until 6am Monday…

I been move/migrating, etc my Hyper-V 2016 VM’s off my HPE hosts, and onto my new Dell Hyper-V 2022 servers…

I used a combination of “Export/Import” for non-critcal VM’s, and as i use Veeam B&R v12, i used that to create a Replication Job to replicate my VM’s onto the new Hosts, then followed the process to Failover, then make permanent Failover.
I made Veeam Backups before that process (all today), and also have the original VM’s on the HPE servers…

All gone OK, apart from 1. The File Server (svr 2016).
It has C: & D: drives (x2 VHDX), and these would have been created in 2017/2018 )
Unsure how, but after a few reboots to make sure all was OK with the VM File server on the new Dell hosts, the D: (Data) vhdx file has become corrupted :frowning:

No idea how, and i dont get any messages to say its corrupted. When its attached to the File server, the server hangs so i cant even login to it :frowning:
I removed the VHDX from the VM, and then mounted the VHDX file on another VM (non-production VM).
I can see the contents, but to click into the folders, its beyond slow, super slow egg timer :frowning:
I have tried to run CHKDSK /F/R against the VHDX in this VM, but after 10 mins, i got the question of “dismounting first” , which i said “Y” too…
But it not moved since (30 mins).

Safe to say this is done for ?

I can bring the original File server VM back online on its original host (HPE server), as there will have been no file changes made (i booted the VM without network just to make sure its still OK which is it)…
Am also doing a Restore using Veeam of the File server (although snapshot is from 1pm, compared to the Replication Migration i did at 7pm) , guess i got that as a backup backup as such…

Moving forward, what would u do ? (i got plenty of storage space on multiple devices, run 10GB network).

adrian_ych (adrian_ych) February 11, 2025, 9:11am 2

I would see how big the VM is and if the VM is running properly on its original location ?
If it is working properly, I would first power it off. Then perform a full backup (or reverse increment or incremental with full synthetic sync).

But I still do not understand why there is no VM migration available for your case (BTW, I am a VMware user). Export and import is literally a no-go for me…the last straw is backup & restore of a powered off VM.

I would not bother with CHKDSK as the vHDD (VHDX or VMDK) would show as a large file, CHKDSK would not know the contents or the details of the file ? The part to question is the Export process, copy of exported file to a location and the import process of th file ?

stuartnicols (stuartnicols) February 11, 2025, 12:36pm 3

The VM, is total 4TB.
C: drive .vhdx file is like 200GB , and system boots up OK with this.
D: drive .vhdx is 3.8TB , with 580GB free space

In Hyper-V , rather than Migration, it is: Failover (from a Replica) , followed by Permanent Failover. (which is what i did on the production server, resulting in the d: drive becoming un-responsive… )

No issue with the VM (server 2016 VM) , its just the D: drive .vhdx file that was nailed.

The VM for now, is in its original location, running on the original HPE server, without issues, and all 170 users are happy (i use re-directed folders for Desktop/Documents along with the File Server storage)…

I plan to attempt the same process this weekend coming again, as am working on the scenario it was a bad replication that got created (i made a new Replica on the new Dell Server) …

adrian_ych (adrian_ych) February 12, 2025, 4:13am 4

What software are you using for VM replication ?
Then I would almost never use Replica from failovers process as a form of migration, unless 100% really necessary and only for powered-off VMs (and only when using VBR).

Coz for me, my critical VMs are using a 3 step protection…

My File Server is only approx 1.4TB, each “subsequent” backup takes like 3 min & the whole Reverse increment Backup process takes like 18 min (including creating the “synthetic full”.

The replication takes approx 7min-12min after the backup process completes.
So I did try a “migration by replication” as I powered off the VM, did a backup…the process took like 28min in total…but note I have existing backups and existing replica running already…