[SOLVED] Backup failed in suspend mode, work on stop mode

[Buzznative]

Hi,

I have a backup problem on Proxmox 6.3, on a hundred VMs backed up daily, I have no errors, for 2 of them, it goes into error in suspend mode, if I do the backup in stop mode manually , no problem.
Bakckup are send to NFS, generate on local tmp folder.

For both, I have this log on backup:

INFO: starting new backup job: vzdump 298 --node [node name] --compress zstd --storage local --mode snapshot --remove 0
INFO: Starting Backup of VM 298 (lxc)
INFO: Backup started at 2021-04-15 09:38:20
INFO: status = running
INFO: CT Name: [lcx name]
INFO: including mount point rootfs ('/') in backup
INFO: mode failure - some volumes do not support snapshots
INFO: trying 'suspend' mode instead
INFO: backup mode: suspend
INFO: ionice priority: 7
INFO: CT Name: [lxc name]
INFO: including mount point rootfs ('/') in backup
INFO: starting first sync /proc/3340/root/ to /tmp/vz/tmp/vzdumptmp1968_298/
ERROR: Backup of VM 298 failed - command 'rsync --stats -h -X -A --numeric-ids -aH --delete --no-whole-file --sparse --one-file-system --relative '--exclude=/tmp/?*' '--exclude=/var/tmp/?*' '--exclude=/var/run/?*.pid' /proc/3340/root//./ /tmp/vz/tmp/vzdumptmp1968_298/' failed: exit code 23
INFO: Failed at 2021-04-15 09:39:03
INFO: Backup job finished with errors
TASK ERROR: job errors

If I launch rsync on node, I have this error for lot of file (certainly for all):

rsync: set_acl: sys_acl_set_file(bin, ACL_TYPE_DEFAULT): Operation not supported (95)
rsync: set_acl: sys_acl_set_file(boot, ACL_TYPE_DEFAULT): Operation not supported (95)
[...]

I don't understand why I have this error on two LXC and for all others, no problem....
On VM, I have no mount point with noacl:

# mount -v | grep acl
#

I don't know what search to resolve this issue, could you help me to resolve?

Thanks

 

[fabian]

your tmpdir does not support ACLs.. set a different tmpdir, or don't use suspend mode.

 

[Buzznative]

In this order, why all others LXC can backup in same mode, same node, same temp dir?

This is a part of backup on last night:

INFO: Backup started at 2021-04-15 03:37:25
INFO: status = running
INFO: CT Name: [lxc1]
INFO: including mount point rootfs ('/') in backup
INFO: mode failure - some volumes do not support snapshots
INFO: trying 'suspend' mode instead
INFO: backup mode: suspend
INFO: ionice priority: 7
INFO: CT Name: [lxc1]
INFO: including mount point rootfs ('/') in backup
INFO: starting first sync /proc/9828/root/ to /var/lib/vz/tmp/vzdumptmp16284_251/
INFO: first sync finished - transferred 2.65G bytes in 418s
INFO: suspending guest
INFO: starting final sync /proc/9828/root/ to /var/lib/vz/tmp/vzdumptmp16284_251/
INFO: final sync finished - transferred 0 bytes in 76s
INFO: resuming guest
INFO: guest is online again after 76 seconds
INFO: creating vzdump archive '/mnt/pve/nfs/dump/vzdump-lxc-251-2021_04_15-03_37_25.tar.zst'
INFO: Total bytes written: 2818447360 (2.7GiB, 3.0MiB/s)
INFO: archive file size: 1009MB
INFO: removing backup 'nfs:backup/vzdump-lxc-251-2021_04_09-22_38_47.tar.zst'
INFO: Finished Backup of VM 251 (00:24:39)
INFO: Backup finished at 2021-04-15 04:02:04
INFO: Starting Backup of VM 252 (lxc)
INFO: Backup started at 2021-04-15 04:02:04
INFO: status = running
INFO: CT Name: [lxc2]
INFO: including mount point rootfs ('/') in backup
INFO: mode failure - some volumes do not support snapshots
INFO: trying 'suspend' mode instead
INFO: backup mode: suspend
INFO: ionice priority: 7
INFO: CT Name: [lxc2]
INFO: including mount point rootfs ('/') in backup
INFO: starting first sync /proc/3070/root/ to /var/lib/vz/tmp/vzdumptmp16284_252/
ERROR: Backup of VM 252 failed - command 'rsync --stats -h -X -A --numeric-ids -aH --delete --no-whole-file --sparse --one-file-system --relative '--exclude=/tmp/?*' '--exclude=/var/tmp/?*' '--exclude=/var/run/?*.pid' /proc/3070/root//./ /var/lib/vz/tmp/vzdumptmp16284_252/' failed: exit code 23
INFO: Failed at 2021-04-15 04:40:05

Effectivly, I can backup in stop mode, but I don't understand the difference between LXC....

 

[fabian]

because those containers have ACLs set on some of their files/dirs, and the others probably don't..

 

[Buzznative]

because those containers have ACLs set on some of their files/dirs, and the others probably don't..

Ok, effectively, I have files with ACL set but acl package not install on LXC...

After install package and remove all ACL (not use in my use case), backup is now OK, thanks for your help !