Can pacemaker retry monitor action on error?

Question

I created simple pacemaker agent for managing MySQL. It works quite well, but there is 1 thing that I can't understand.

When resource is "Started", and monitor action returns one of:

$OCF_NOT_RUNNING
$OCF_ERR_GENERIC

pacemaker immediately stops resource, and run it again. Is there any way to do more than 1 monitor check, before stopping resource? E.g. 5 retries in 30 seconds timeout. Of course I can set operation timeout, but that's all.

I know that I can implement retry mechanism inside of my resource agent, but I don't feel that it's the best approach.

Any help will be appreciated

Hello @FedericoSierra, thanks for your comment. I put it on github, so you can check metadata https://github.com/TomaszUrugOlszewski/pacemaker-mysql-debian — Tomasz Olszewski, May 15 '15 at 16:20
I temporarily added retry to script, but my question is still open :-) — Tomasz Olszewski, May 16 '15 at 18:30
Tomasz I'm seeing your code. Try to use `migration-threshold` greater than 1 — Federico Sierra, May 16 '15 at 19:03

score -1 · Answer 1 · answered Apr 02 '19 at 13:13

-1

you have to implement retry mechanism, otherwise you will not be able to retry monitor action error.

answered Apr 02 '19 at 13:13

marlin putnam

11
1

It might be worth expanding your answer a little, as it is prone to misinterpretation. – womble Apr 03 '19 at 01:54

Can pacemaker retry monitor action on error?

1 Answers1