"holding" backup and snapshots for a while

[greg]

Greetings

I have a cluster of servers which are regularly doing backups and snapshots.

During some production-critical periods, I need to "hold" such traffic/load which could impact the performances.

What do you think would be the proper way to do that?

Thanks in advance

Regards

 

[fabian]

for backups, you could implement this using a hookscript that checks some condition and fails the backup task(s) or the whole job. for snapshots no such option exists I think, but snapshots should be almost free except for the "ensure consistency" freeze part if multiple volumes are involved.

 

[greg]

Thanks for your answer.
Maybe I could use

pct set --lock

? Then all operation would failed or be postponed?

 

[fabian]

you could do that as well, but it blocks *all* tasks, including things like HA migrations, ... the backup storage hookscript variant is more targeted and easier to control

 

[greg]

Yes blocking all tasks in what I want. I definitely don't want a migration in the middle of a live event What would be the preferred parameter? I tried "mounted" and it said:
CT is locked (snapshot-delete)

 

[fabian]

there isn't a custom lock value yet, and some of them have special semantics and allow being cleared/ignored for certain tasks.

in your case, it seems the container was already locked (a still ongoing, or failed snapshot deletion?)