Master, commit a3139 failing backups

[manilx]

@simonp NFS and S3

[marcoi]

I just got hit with the backup error, not sure what version i was running but it was likely a3139 or higher. I have XO rebuild each morning to fetch latest updates.
Error I can find (interface jumped to v6 so still learning)

Can't init vhd directory without using alias

"Error: Can't init vhd directory without using alias\n    at RemoteVhdDisk.init (file:///opt/xo/xo-builds/xen-orchestra-202602170700/@xen-orchestra/backups/disks/RemoteVhdDisk.mjs:78:31)\n    at async _mergeVhdChain (file:///opt/xo/xo-builds/xen-orchestra-202602170700/@xen-orchestra/backups/_cleanVm.mjs:75:5)"

"Error: Can't init vhd directory without using alias\n    at RemoteVhdDisk.init (file:///opt/xo/xo-builds/xen-orchestra-202602170700/@xen-orchestra/backups/disks/RemoteVhdDisk.mjs:78:31)\n    at async _mergeVhdChain (file:///opt/xo/xo-builds/xen-orchestra-202602170700/@xen-orchestra/backups/_cleanVm.mjs:75:5)"

Backup to NFS share

[simonp]

@manilx @acebmxer @marcoi

Hi, thanks for the reports, we were able to fix an issue with the health check but we weren't able to reproduce the backup issue.

Could you checkout the "debug_backups" branch, try to run the backups again and send the logs here ?

There should be no need to restart or rebuild.

Thanks !

[marcoi]

@simonp I downloaded the branch but i have no idea how to run a backup in v6 interface and v5 isnt working with the latest updates to XO. How do i run a backup job manually in v6?

[florent]

@marcoi can you go to <xo ip>/v5/ ?

[marcoi]

@florent I posted this in the v6 master thread, but it opens to the old interface but sits with a still loading message.

[acebmxer]

@simonp

Updating now and will report back

XO-CE loaded just fine for me.

Just to confirm correct commit is - 29b7a backup failed log attached.

2026-02-17T14_25_39.007Z - backup NG.txt

[simonp]

@acebmxer

Thanks a lot for the logs, we've been able to identify a possible cause for this issue, could you pull the "debug_backups" branch and try running the backups again ?

Hash is cb85e44ae

[acebmxer]

@simonp

[info] Updating Xen Orchestra from '29b7a67dc' to 'cb85e44ae'

2026-02-17T15_25_36.261Z - backup NG.txt

[marcoi]

I got back into v5 url once i removed LISTEN_ADDRESS="" from my xo-install.cfg

the backup ran on Xen Orchestra, commit cb85e

[simonp]

@acebmxer

Thanks a lot, the first issue was resolved but your log indicates a separate issue.
We think we resolved it on the "debug_backups" branch, could you test it again ?

Anyways, if this fixes the backup issue, a fix should be available tomorrow Paris time on master.
In the mean time you can revert to the commit a3139 as explained in the first post.

Thanks again for your help.

[acebmxer]

@simonp

[info] Updating Xen Orchestra from 'cb85e44ae' to 'eed3d72f7'

First log is from automatic schedule before latest update.
2026-02-17T18_00_00.012Z - backup NG.txt

latest log after update. Backup completed successfully.
2026-02-17T18_22_52.199Z - backup NG.txt

[acebmxer]

Updated to commit -db8ad

Backup successful...

2026-02-18T14_09_09.531Z - backup NG.txt

[manilx]

Confirmed. Backups run successfully again.

[marcoi]

Updated to commit -db8ad this morning and backup is working.

[simonp]

Thanks for the heads-up and thanks for the help !

[manilx]

I've just been hit with this also: Can't init vhd directory without using alias

All S3 repo's running SeaweedFS, where fine the last time I used them (a few weeks ago). I don't use my xcpng servers on the homelab daily as I move to Unraid.

Running Master, commit 99312

[jr-m4]

@manilx said:

I've just been hit with this also: Can't init vhd directory without using alias

All S3 repo's running SeaweedFS, where fine the last time I used them (a few weeks ago). I don't use my xcpng servers on the homelab daily as I move to Unraid.

Running Master, commit 99312

I'm seeing the same error, with the same reported commit.

[pierrebrunet]

@jr-m4 Hi,
We have found this issue in another thread, it is now merged