hi guys,
I would like to create an endless action (policy) that will reapply itself endless times, even if it failed endless times.
I’ll explain why:
if my computer relevant to the action but fails because of temporary problem and it took me some time to fix the issue on the specific computer, it may be that I fixed it after the failure retry reached it’s maximum retries.
you can see the screenshot of my action:
I’ve also tried the “on failure, retry > wait 15 minutes between attempts”, but still of course it reaches the limit of retries
is there any option to retry the action endless times depend less if it failed?
In general, the approach would be to create a “relevance” that detecs the existance of a specific applicability condition, and then use the applicability option “Whenever it becomes relevant again”.
In your specific scenario it appears that you have to manully fix “temporary” problems. In such case your “relevance” should ideally check those temporary problems and becomes relevant only after they have been fixed
One possible option is to change the success criteria. If you are using a Fixlet, the default success criteria is that the original relevance evaluates to False. If you change that to “…all lines of the action script have completed successfully” that should prevent the action from being seen as failed and avoid the retry limit you are hitting.
Maybe the ideal option though, depending on your use case, is to create a fixlet to detect the temporary state that impedes the fixlet from completing successfully so you fix the issue impeding your original fix as well as the original issue.
Or, have a “relevance killer” statement, i.e., the relevance is something like “not exists /var/opt/BESClient/ignore/site-content-XXXX.lock” where you can substitute ‘content’ for fixlet, task, or whatever - where XXXX is the ID of the content. Adding ‘site’ label is, of course, optional.
When you need to stop the policy (because you are working on it) you need to merely touch a file with the correct name. And remove it to restart the policy. More automation is always possible.
