-
Notifications
You must be signed in to change notification settings - Fork 82
Description
Apologies if an 'issue' is the wrong place for this. I couldn't find reference to a spot for Q&A.
I'd like to preface this by saying I realize I've got a weird little edge case here, and I'm not expecting there to be a solution, but perhaps I'll get lucky and someone has an idea. Or maybe it's not as weird an edge case as I think, and it's worth a feature request.
So, my question: Is there a way to get needrestart to ignore certain processes based on some criteria? I'm thinking cgroup, or GID, or a list of UIDs, or anything along those lines.
I've had a look through the code; I don't see any obvious solution to this, but my perl is a bit rusty and so perhaps I've missed something.
My use case:
I run some shell servers with researchers on them. Sometimes they're running their own compiled code for long periods of data analysis, and can't conveniently stop those processes for routine shared library updates. Sometimes someone's doing something even more out of the ordinary, like the gdb process that has been running for the last 24 hours on a now-deleted program, and the person who started it isn't easily reachable.
I'm not too concerned with these problems, and I'd like to be able to exclude them from the Nagios-mode run of needrestart that is in my monitoring, and instead just have something report on those to me less frequently, and less urgently.