Error Duplicating VHDX files

[crembz]

Hi there,

 

I'm in day 3 of evaluating drivepool after a horror experience with migrating a storage space from windows 10 to server 2012r2 which has cost me a significant amount of $$$ for recovery software and spare HDDs.

 

I used drivebender about 4 years ago on WHS and thought it was quite an elegant solution. I see drivepool seems to have come a long way and I love the fact it's file based ... not complicated recovery process.

 

I have run into an issue where when I try and replicate a folder containing live Hyper-V VMs I get an error about oplocks and permission. I've attached a screenshot.

 

If I turn the VMs off, duplication continues, however hyper-v loses access to the VM config and VHD files.

 

Is this a known issue and is there a way around it?

 

Also, how reliable is snapraid when used with drivepool? I love the idea of mirroring important folders, but for 90% of my files, mirror is overkill. I'm also using drivescanner and was wondering whether this is a sufficient protection mechanism against drive failure?

 

Thanks.

 

 

[Christopher (Drashna)]

The file is in use.

Specifically, the "oplock request" stuff explicitly refers to file locks. 

 

 

As for "lose access", I'm not exactly sure what you mean by that. Do you mean that HyperV doesn't see the files when this is happening? If so, that's very unusual. While duplicating, the original files are left in place, but we do create ".copytemp" files on the destination drives until the duplication process is finished. 

 

But it may be worth letting the duplication finish, (like completely overnight), and then see how it goes. 

 

Otherwise, please open a ticket here: https://stablebit.com/Contact

 

 

 

 

As for SnapRAID:

http://community.covecube.com/index.php?/topic/52-faq-parity-and-duplication-and-drivepool/

Specifically, we don't actually support it being used with our software. 

However, that doesn't mean that it wouldn't work. 

 

 

But you should definitely read the first part. 

[crembz]

Thanks for the response.

 

I need to clarify the issue with the files being unavailable. I currently have the VMs in a non-duplicated folder on the share and they are running. Once I turn duplication on, the errors present themselves. If I turn the VMs off, and then enable duplication the process starts and Hyper-V reports problems reaching the files. This also happens if I remove a drive from the pool while the re-balancing is occurring. I'm not sure what happens during re-balancing though I suspect the active files are not rebalanced.

 

I've attached the error which Hyper-V throws. This screenshot was taken whilst removing a drive (same error as when duplicating).

 

One other question about space savings ... I'm debating which of the following would prove more effective:

 

Drivepool duplication + folder compression

Drivepool no duplication + parity (snapraid)

 

I'm leaning towards the first option as being simpler and more supportable.

 

Oh and I just wanted to say I think this is a great product. Very pleased with the ease and performance I'm getting thus far.

[crembz]

OK just did some more testing and compared this to my drive bender setup. Seems drive bender also has issues working with running Hyper-V VM files. at least drive pool lets me know what the problem is, drive bender just sits there infinitely processing.

 

I've also noticed that migrating storage of the drivepool drive, throws an error in hyper-v (attached). This can be avoided by turning the VM off before doing a storage migration.

 

 

[Christopher (Drashna)]

I think I responded to a ticket about this.

 

As for the migration, I'm not sure why that would be occurring. If you could, grab the "System" and "Applications" sections of the "Windows Logs" in the Event Viewer of the hyper-V host (run "eventvwr.msc")

[crembz]

Thanks Chris, Yes I got the response. It all seems related to the locking that occurs when you perform certain operations like removing a drive. I wasn't aware that the pool goes into read-only mode. I also wasn't aware that it can not duplicate files that are in use.

 

For a 24x7 VM, I understand that the pool will duplicate the file on creation. How are updates to the file duplicated while it is running / open?

 

Oh and I tried a storage spaces migration from server 2016 to windows 10 which *should* be using the same storage spaces version ... no luck, the ReFS mirrored volumes were presented as 'unformatted' in win 10. My drivepool pool was ... perfect

[Christopher (Drashna)]

Yes, the software does put the pool into a read only mode while removing disks. This is in part to prevent new locks from being created, so you can successfully remove the drive. And to prevent new data from being added and causing issues.

 

 

As for the duplication:

When duplicating the files, the software copies the file to a ".copytemp" file on the destination drive. Once that's finished, we check to make sure it copied properly, and then rename the files. It's then "duplicated" and updates the statistics. Also, we actually use the normal Windows API for copying the file (basically the same as what Explorer uses). And we can't copy locked files, because they may be updated after the copy starts... causing corruption, or worse.

 

So on open files, we just skip them and flag them for the user, so they know that there is an issue (as you've seen).  

 

But once the files have been duplicated, any modifications will be done to both copies of the files. So once it's duplicated, you should be completely fine (and work just as well as newly created VMs). 

 

 

 

 

And ouch, on the Storage Spaces thing. Though, it's slightly different versions, but still. That's ... painful at best.