Backup error with space available

Bane

@olivierlambert I'm also trying to restore the VM's that were successfully backed up. I have two that fail to restore and give the error:

IMPORT_ERROR_PREMATURE_EOF()

What can I do to determine what this failure means? I found a previous entry about disk space, but I do not have any disk usage issues that I can see.

Filesystem          Size  Used Avail Use% Mounted on
devtmpfs            2.8G   12K  2.8G   1% /dev
tmpfs               2.8G  268K  2.8G   1% /dev/shm
tmpfs               2.8G   10M  2.8G   1% /run
tmpfs               2.8G     0  2.8G   0% /sys/fs/cgroup
/dev/nvme0n1p1       18G  1.9G   15G  12% /
xenstore            2.8G     0  2.8G   0% /var/lib/xenstored
/dev/nvme0n1p5      3.9G   83M  3.6G   3% /var/log
/dev/nvme0n1p4      512M  3.0M  509M   1% /boot/efi
tmpfs               571M     0  571M   0% /run/user/0
tank                4.3T  256K  4.3T   1% /tank
tank/local-store    4.3T   17G  4.3T   1% /tank/local-store
//192.168.2.15/iso   22T   15T  6.5T  70% /run/sr-mount/39ac011c-a02a-d9bb-0fef-7fffebbefacc

Thanks

olivierlambert

It's not a fix, it's just a workaround. It's not normal to have this kind of timeout
IMPORT_ERROR_PREMATURE_EOF() means the backup is maybe incomplete or import stalled

Bane

@olivierlambert

Is there anyway to determine whether it's a failed backup or stalled import? Is there anything that can be done to correct a stalled import? I assume there is nothing that can be done for it if it's a failed backup..

Also interesting is that the restores that fail are listed with a number instead of the VM name.

xcp-ng restore error.png

olivierlambert

It's really hard to tell without being inside the infrastructure

Bane

@olivierlambert

Any logs inside XCP-NG that I can look at? I've resigned myself to the fact that I will have to rebuild the two VM's that won't import, but I'd like to prevent a similar situation in the future.

olivierlambert

Are you using XOA or XO from the sources? In any case, be sure you be entirely up to date (on latest or latest commit)

Bane

@olivierlambert

I have both installed to try out community. I am performing the restore from XOA, the same console I used for the backups.

Incidentally, I would pay for the XOA, but $70 a month to get back up for my home lab is a bit pricy, I only have one server. You guys might consider offering a paid version with backup that maybe doesn't include premium support for a reduced price. I can't speak for anyone else, but I would be willing to pay for something like that, to have a turnkey solution with easy updates and backup.

olivierlambert

XOA isn't meant for individuals but only companies We can go below $70 because it's not even profitable at this price already.

edit: when I said try with XOA, I'm not suggesting to pay for it! Just go for the trial, to see if it works there. If it is, it might tell that's a build problem on your side

Bane

@olivierlambert

XOA is what I have been trying. I have not tried the restore in community.

olivierlambert

So the backup was also made with XOA in the first place?

Bane

@olivierlambert

That is correct.

olivierlambert

You should probably open a support ticket with a support tunnel so we can take a look remotely.

Bane

@olivierlambert My trial of premium is over, so I am unable to do that.

olivierlambert

You should be able to do it anyway

Bane

@olivierlambert I've opened a tunnel. The ID is 33769. Do you need any other information from me?

olivierlambert

That would be better if you can open a support ticket from xen-orchestra.com website, so the support team could take a look tomorrow (we are UTC+2 here)

Bane

@olivierlambert After the latest update, one of the two VM;s that would not restore now restores successfully. I'm still not sure why the second one won't restore.

olivierlambert

That's really weird