[SOLVED] Backup of containers are failing when they were restarted before

[ojaksch]

Have this observed a few times while on PBS beta but did a fresh test with PBS stable yesterday.
Most of the containers are Debian Buster from Proxmox' template. Did some updates to some of them and restarted the concerned ones by systemctl reboot, like I always do. Every CT I did this with failed to backup last night.

This is a very strange behaviour. CT's that are restarted by systemctl are working as normal, but are failing to the next backup. When I remember correctly the backup after next (tonight) will be working as normal.

• Edit #1: Did no update nor a reboot to any CT yesterday and backups went fine.
  Next test will be an update and reboot to one CT and additionally an update and shutdown to another CT.
• Edit #2: I had rebooted CT110 and shutdown/start CT100, but this night something went completely wrong. Untouched CT101 also failed to be backuped, CT102 and CT104 went fine, all other CT's are seen as "todo", but backup job never finished and canceled after CT104:
  
  

  VMID STATUS TIME SIZE FILENAME 100 err 00:02:08 0.00MB -
  101 err 00:02:35 0.00MB -
  102 ok 00:09:54 0KB ct/102/2020-11-20T23:19:45Z
  104 ok 00:01:57 0KB ct/104/2020-11-20T23:29:39Z
  108 todo 00:00:00 0.00MB -
  109 todo 00:00:00 0.00MB -
  110 todo 00:00:00 0.00MB -
  111 todo 00:00:00 0.00MB -
  116 todo 00:00:00 0.00MB -
  118 todo 00:00:00 0.00MB -

  
  While analyzing this I discovered that there are some of my bash-ish proxmox-backup tasks still running, which are collectiing and backup'ing tons of user data and other stuff while beginning of PVE's backup job. I wager my grandmother's ass that they are getting in each other's way, as the PBS host is a small, elder and weakly hardware. I'll start PVE's backup job at later time next night.
• Edit #3: Nothing new - While reorganizing the times of backups I choosed the wrong backup space for the CT's. Better luck next time.
• Edit #4: I displaced the start of PVE's backups to a later time - a time after a dozen of proxmox-backup-client bash jobs are definitly finished - and since then everything is working as normal. Weird. Especially weird as only the rebooted CT's failed, as described at threads beginning. Whatever, I'm sure that's because of the powerless hardware I'm using for the PBS server.

Problem recognized, problem solved.

 

[wolfgang]

Hi,

are you Proxmox VE also updated. Because it is important that the Proxmox VE site has also the last updates.

 

[ojaksch]

Jein Last update on PVE and PBS ran at 2020-11-11 / 14:27, just a short time after the release of PBS v1.0.1
I'll run my next test as described before and after that I'll update both hosts and recheck.

Edit: No or no relevant updates available.