[SOLVED] Multiple backups parallel since 6.3 - hanging system

V

voarsh

Member
Nov 20, 2020
218
19
23
28
Since around 6.3 newer updates, Proxmox has been running multiple backups, some on the same disk, not all, and I've had 3 at the same time - previously it did one at a time and didn't start new ones until the previous one completed.

Today my system broke badly, I managed to stop the tasks but I/O wait has sent my load to 300, Proxmox doesn't load but my applications are running slowly.

Anyone know about how to get back the old backup queue behaviour? My backup jobs can run for 1-3 hours and I really need the queue behaviour back.

Not much luck with PBS product, so I am stuck with native Proxmox backups.
 
Hi,
this was a regression that has been fixed in pve-manager >= 6.3-7. Please upgrade and see if it works.
 
  • Like
Reactions: voarsh
Fabian_E said:
Hi,
this was a regression that has been fixed in pve-manager >= 6.3-7. Please upgrade and see if it works.
Click to expand...
Solved it.
I however, run into another issue where the backup waits to start for 2/3 hours and timesout. Seems disks are bigger and much slower to complete (without PBS) so many are failing now.

Anyway to increase the timeout?
 
voarsh said:
Solved it.
I however, run into another issue where the backup waits to start for 2/3 hours and timesout. Seems disks are bigger and much slower to complete (without PBS) so many are failing now.

Anyway to increase the timeout?
Click to expand...

Yes, there is a vzdump CLI option for this:
Code: 
       --lockwait <integer> (0 - N) (default = 180)
           Maximal time to wait for the global lock (minutes).
 
Fabian_E said:
Yes, there is a vzdump CLI option for this:
Code: 
       --lockwait <integer> (0 - N) (default = 180)
           Maximal time to wait for the global lock (minutes).
Click to expand...
Can you clarify if that is for each backup task, ran via cli manually?
Or can this be set globally for all backups, that happen when scheduled?
 
To set it for a single job, edit /etc/pve/vzdump.cron and add it as a CLI option. To override the default for all jobs on a given node, edit /etc/vzdump.conf.
 
  • Like
Reactions: voarsh
Fabian_E said:
/etc/pve/vzdump.cron
Click to expand...
FIY, thanks so much for this tidit of info. I was feared backup schedules were stored in a .db file or something, that would be cumbersome to edit by hand. This cron style format makes it pretty easy to download/manipulate and place it back - especially when you have 50+ backup jobs with different storages! :)
 
You must log in or register to reply here.
Top Bottom