Create Backup, not enough local space

[0C0C0C]

I'm trying to move a big Container from an Debian 9 Proxmox 5 System to a new Debian 10 Proxmox 6 System. The Container has a local disk on ZFS and 2 Mount Points on ZFS.

When i'm trying to create Backup,i got the problem that i do not have enough local space.
I mounted an external Storage using SMB, but when im trying to create a backup to this storage, Proxmox still forces to use the local storage to create the backup and its not working because... not enough space.

I tried to find Solutions here and using google but the only thing i can find is: use backup and restore but nothing handling my problem.

The old System VM Storage is a manual created ZFS Pool, on the new system i got 2 ZFS Mirrors created in Proxmox 6.

The only idea i found is, creating the dump using a shell, and define the tmpdir in the vzdump command, but i dont know if this is working because... proxmox already forces not using a direct backup to my SMB Storage and takes a local TMP Directory.

 

[dietmar]

Maybe you forgot to allow backups on that storage? Just set content type 'backup' for that storage...

 

[0C0C0C]

No, it's possible to chose the CIFS Storage as Backup Target, but when starting a backup, this gets shown:



Proxmox uses the local storage (/var/tmp/...) for temporary files, and the local storage has not enough space to create the backup.

What i tried now, was changing the /etc/vzdump.conf and change the tmpdir var to a folder in my external Storage mount. The external Backup Storage is now mounted using SSHFS instead of CIFS and added to proxmox as "Directory". When starting a backup now, i get these errors:



But... the Backup is running... i just see a *.dat file instead of a *.tar file, but maybe because the backup is still running. The SSHFS Mount is done using options "defaults,_netdev".

 

[0C0C0C]

OK backup is done now, but restore on new server fails...

extracting archive '/opt/sh2tosh3/dump/vzdump-lxc-107-2019_07_27-12_14_28.tar'
tar: -: Cannot open: Permission denied
tar: Error is not recoverable: exiting now
TASK ERROR: unable to restore CT 101 - command 'lxc-usernsexec -m u:0:100000:65536 -m g:0:100000:65536 -- tar xpf - --totals --one-file-system -p --sparse --numeric-owner --acls --xattrs '--xattrs-include=user.*' '--xattrs-include=security.capability' '--warning=no-file-ignored' '--warning=no-xattr-write' -C /var/lib/lxc/101/rootfs --skip-old-files --anchored --exclude './dev/*'' failed: exit code 2

After Moving the dump to the local storage of the new server, it starts, but again with 2 Errors, like when creating the backup, but: the restore is running right now... soooo, maybe the Container works after it, we will see...

extracting archive '/stor-zfs-2tb/test/dump/vzdump-lxc-107-2019_07_27-12_14_28.tar'
tar: ./etc/vzdump/pct.conf: Cannot change ownership to uid 208687, gid 208687: Invalid argument
tar: ./etc/vzdump/pct.fw: Cannot change ownership to uid 208687, gid 208687: Invalid argument

 

[0C0C0C]

OK it fails completely...

tar: ./etc/vzdump/pct.conf: Cannot change ownership to uid 208687, gid 208687: Invalid argument
tar: ./etc/vzdump/pct.fw: Cannot change ownership to uid 208687, gid 208687: Invalid argument
tar: ./var/spool/postfix/dev/urandom: Cannot mknod: Operation not permitted
tar: ./var/spool/postfix/dev/random: Cannot mknod: Operation not permitted
Total bytes read: 209769226240 (196GiB, 674MiB/s)
tar: Exiting with failure status due to previous errors
TASK ERROR: unable to restore CT 101 - command 'set -o pipefail && cstream -t 0 | lxc-usernsexec -m u:0:100000:65536 -m g:0:100000:65536 -- tar xpf - --totals --one-file-system -p --sparse --numeric-owner --acls --xattrs '--xattrs-include=user.*' '--xattrs-include=security.capability' '--warning=no-file-ignored' '--warning=no-xattr-write' -C /var/lib/lxc/101/rootfs --skip-old-files --anchored --exclude './dev/*'' failed: exit code 2

 

[Stoiko Ivanov]

Try restoring as a 'privileged' container (the default in the GUI got changed to unprivileged recently - and the last error you posted indicate that this used to be a privileged container.

hope this helps!

 

[0C0C0C]

Ah shit lol you're right... it worked... now i need to test if a restore in shell using "-unprivileged" with "-ignore-unpack-errors 1" is working to get an unprivileged container of the privileged backup.

 

[Stoiko Ivanov]

hmm - not sure the restore and lxc start do quite a few things in between - so I never considered this an option.
Please report how it works out for you!

an alternative could be to setup a new unprivileged container and use `rsync` from within the containers to get all relevant data to the new container and then switch the config (name, ip, ....)

Good luck!

 

[0C0C0C]

I got expected errors, but the Container was created and not destroyed like in GUI Restore, and it is starting, but because of the used Services inside of the VM i get some problems with mounts and services.

extracting archive '/stor-zfs-2tb/test/dump/vzdump-lxc-107-2019_07_27-12_14_28.tar'
tar: ./etc/vzdump/pct.conf: Cannot change ownership to uid 208687, gid 208687: Invalid argument
tar: ./etc/vzdump/pct.fw: Cannot change ownership to uid 208687, gid 208687: Invalid argument
tar: ./var/spool/postfix/dev/urandom: Cannot mknod: Operation not permitted
tar: ./var/spool/postfix/dev/random: Cannot mknod: Operation not permitted
Total bytes read: 209769226240 (196GiB, 648MiB/s)
tar: Exiting with failure status due to previous errors
Detected container architecture: amd64

So best and safe way is really to create new containers and reinstall the services / transfer the data.