VHD export and (delta) backup stuck at 99%
-
Hi,
I am having a bit of a strange issue;
XCP-NG, current release, XO build from source, latest build.
The problem:
When making a backup or exporting the disk to a VHD, the export in XCP-NG will get stuck at 99%. The backup target will receive the VHD and data is transmitted. After a period of time the created VHD will be removed and XO will give up. The task will stay at 99% in XCP-NG, only be removed after restarting the toolstack.This is how the job looks in XO:
The HTTP connection has timed out seems to be the connection between XCP-NG and XO (XO is a VM on this XCP-NG machine), there is no sign of actually being a network related issue.
What does, and what does not work:
5 VM's which were physical machines and made virtual by using microsoft disk2vhd (Windows 2019 machins). All of these will get stuck at 99%, tools are installed and running, and the vm's perform like expected otherwise.2 VM's which were migrated from Hyper-V (Windows 2022 and Debian 11) do succesfully backup.
1 VM made from scratch on XCP-NG also work (debian 11).
What I tried:
Other backup target (local, fast), no difference.
Converting the original .vhd file in hyper-v to dynamic vhd (was already a dynamic vhd but I figuered I can at least try), also no difference.Any suggestions?
-
Any errors shown in the browser console? Have you checked the logs for additional details / clues?
Did this ever work correctly with these VMs?
-
Hi,
Can you provide the commit number you are building XO from?
-
@olivierlambert Errors on 79671e6d6, I am now trying commit 5c0b29c51
-
@Danp Has never worked for these VM's, I will look into the logs. Is daemon.log expected to have information about this?
-
What's your NodeJS version?
-
root@xo:~# node -v v18.13.0 -
Trying in latest commit, no change in result.
This is what the task view looks like couple seconds before it thinks it is ready:
.vhd backup file in SMB target (test target)
Then, couple minutes later, it just seems to give up:
the exported .vhd is also gone:
result of job:
log attached
-
I would try to migrate or to full clone the disks of the "broken" VMs. And see if exports after that (after XCP-ng will recreate entirely the new VHD from scratch)
-
@olivierlambert I tried a full clone, which also remains stuck at 99%, I can try a clone in XCP-NG Center maybe? I currently have nothing else to migrate to from this machine.
-
@OhSoNoob Does it also hang on an XVA export?
-
So the issue might be on the VM disk directly, and having the built-in XCP-ng VHD tools having issues to export it
-
Full clone/copy via XCP-NG center indeed stuck at 100% (not 99%):
But in XCP-NG center there comes a warning up, result out of range, can this be a clue?
@Danp let me try an XVA export next
-
@Danp Exporting to XVA (no compression) does work!
After importing the exported machine, the export of the VHD works on that machine!
I am now doing the same for the 4 other machines to get a 5/5 score.
By the looks of it, this seems to be the solution!
BTW, uploading de XVA in XO didn't work for some reason, it prompts about prepairing and then just stops. Importing via XCP-NG center does work fine however.
-
I think it's because the XVA format is different (a bit weird but not the point here). So it helps to recreate something different that can be then re-imported, and solve the problematic original VHD.
It's a funny workaround, but glad it worked
-
@OhSoNoob said in VHD export and (delta) backup stuck at 99%:
uploading de XVA in XO didn't work for some reason, it prompts about prepairing and then just stops.
Hmmm.... never encountered that situation.
