ESXI 6.7 -> Hyper V Question(s)

[JMattinglyIT]

I have an old ESXI server in Ohio (I am in Atlanta). I am currently checking out Starwind as the tool seems really simple to use compared to Windows Admin Center. I had a few quick questions:
1. I've got a test VM I've backed up to Atlanta. When I load this VM into Hyper V and attempt to boot (pointing a new VM to the newly created Hard Disk Image) it fails to boot and moves into recovery. I've tried startup repair but that has failed. Am I missing a step here?

Regarding my Starwind version and the conversion process I am going through:

I am running version 9.0.1.772
The test VM is currently powered down

When I check the logs the only thing I do note appears to be tied to timeouts with the ESXI host, but the conversion appears to succeed regardless.
As a precaution I've extended the timeout parameters for my TCP connections- if I don't get a reply in here I'll give the whole process another go from the beginning once my host has rebooted and restarted services.

[yaroslav (staff)]

Welcome to StarWind forum! The problem you explained relates to drivers, not V2V Converter.
Please make sure to delete anything that relates to VMware tools before conversion.

[JMattinglyIT]

Got it- so the procedure should be

1. On Host Vcenter VM: Uninstall VCenter tools, allow it to cleanup and probably reboot
2. On Destination server, run migration on the target VM
3. Following migration, run startup repair to ensure system boots w.o issues

I had one additional question regarding a SQL environment specifically

I have a SQL server with four drives (Including C:).

I attempted a migration on this machine earlier in the week but did not consider the VMware tools factor- I noticed that the migration, once complete, had compressed all four unique drives into a single virtual hard drive.

Would it be better to target each drive individually and migrate them to a VM in my hyperV environment, choosing to attach each drive after the migration?

[yaroslav (staff)]

Hi,

Thanks for your update.
That's strange. Are those volumes physically on one VHD(x)?

Would it be better to target each drive individually and migrate them to a VM in my hyperV environment, choosing to attach each drive after the migration?

Yes, please.

[JMattinglyIT]

Each drive is independent from what I can see in VCenter. It definitely is strange but I am fine with moving each drive during the course of the day.

Thank you and the Starwind team for creating such a useful tool and being so responsive on the forums.

[yaroslav (staff)]

Thanks for your kind words and sharing something interesting
Have a nice weekend.

[rajdude]

Each drive is independent from what I can see in VCenter. It definitely is strange but I am fine with moving each drive during the course of the day.

Thank you and the Starwind team for creating such a useful tool and being so responsive on the forums.

I have successfully migrated hundreds of VMs using Version 9.0.1.772. Some of them are large SQL servers well over 10 TB and having over 10 drives in VMware.
Use that version only!
My tips:

• No need to separately migrate the disks.
• Make sure you uninstall the VMware tools and
• shut down the VM BEFORE running V2V tool.
• Make sure there are no snapshots in VMware.

[JMattinglyIT]

So the VM absolutely must be offline?

I only ask because this environment only has the one DB server and it will be challenging to find a time when this box can be completely shut down due to the systems that rely on it. Probably my best bet would be late night or over the holiday but even still it could result in data loss.

- I'm running 9.0.1.848 as my starwind v2v converter currently.

So far I've uninstalled Vmware tools on my target machine and rebooted it, currently I'm analyzing the logs because I am seeing something odd

12/19 16:18:52.725 c38 Facility ???INFO: WrapperESXI::Wait: Progress - 25
12/19 16:18:53.256 c38 Facility ???INFO: CPPConverting::ConvertDisks: pathInput: [Raid5] myserver/myserver-000010.vmdk, pathOutput: D:\\myserver\myserver.vhdx
12/19 16:18:53.335 c38 Facility ???INFO: ESXVmdk::ParseFile: sFilename - [Raid5] my-server\my-serverpath.vmdk
12/19 16:18:53.335 c38 Facility ???INFO: ESXVmdk::ParseFile: m_cvmxSpec - moref=14
12/19 16:18:53.428 c38 Facility ???INFO: ESXVmdk::ParseFile: sslThumbprint - FE:4C:9F:BC:EE:2D:82:1B:F2:68:AC:20:B9:26:3D:8C:2E:AF:9A:38
12/19 16:18:56.053 c38 Facility ???INFO: CPPConverting::ConvertDisks: CreateInputImage: 0
12/19 16:18:57.897 c38 Facility ???INFO: CPPConverting::ConvertDisks: CreateOutputImage: 0
12/19 16:21:01.823 2fe8 Facility ???INFO: VMConverter::ESXIConverter::Connection::<lambda_4eab79f5f0516eba86b7810fe0d8b684>::operator (): Timeout ESX connection. Performing updates.

This is what I see- the line indicating timeouts extends onward for a few hours, then the application closes itself. I've successfully migrated other machines using Starwind onto the same destination and was still receiving timeouts, but I'm trying to figure out if this has an impact on anything. When I check the destination directory the VHDX is always there so I'm inclined to think I can ignore the timeout.

The past few times I've tried to migrate my SQL instance it appears to complete, but the VM will not boot and cannot repair itself.

[yaroslav (staff)]

VM can be online.

The past few times I've tried to migrate my SQL instance it appears to complete, but the VM will not boot and cannot repair itself.

This issue points to some sort of driver mismatch between the source and destination. Try converting the individual disks and creating the VM out of them at the destination.

[JMattinglyIT]

Yeah this absolutely must be tied to drivers or something

I'm trying one final time, today with a few parameters adjusted. If I can boot from the 100G C:\ drive where the OS is stored I will post in here to let everyone know. I know this server is running windows server 2019 standard, I have a disc file I can try to load as part of a startup repair process as a backup. It's just odd that this single drive is not working. If none of this works I'm going to need to find a weekend to shut the server down and try again.

[yaroslav (staff)]

Please keep me posted.
Also, try Veeam instant recovery or disk2VHD for C:\.

[JMattinglyIT]

No success this morning- cloning C drive results in errors stating it cannot find the OS. Tried the DVD but it failed as well.

Part of the problem is that, and I hadn't thought of this just in my race to get this done, the C drive is part of a RAID array on the VM.

I'll need to see if maybe the Veeam option is valid. Again I have to tread carefully as this is their only SQL environment- its very far from ideal.

[yaroslav (staff)]

Hi,

Try p2v from inside the VM.
AFAIK StarWind V2V and Veeam B&R are non-destructive to the source.

[JMattinglyIT]

Interesting Idea,

So just to clarify that would mean we install Starwind on this server and run a P2V with the destination being the hyperV box?

[yaroslav (staff)]

Yup, that's right.