Jarek Potiuk
Jarek Potiuk
The PR needs to be rebased and conflicts removed - it's actually 149 commits behind main, so rebasing is a good thing.
IF I am not mistaken - this one also adds timeout in supervising process. Was that intentional ? (the commint message only mentions adding SIGHUP/SIGTERM) ?
Hmm. I am not sure @ashb - how much this one will be affected by AIP-72? It's quite likely (almost certain) that starting task will be heavily affected in Airflow...
> Hi @potiuk @ashb ! Can you share please some more details about AIP-72? As i can see, there was a page dedicated to some proposals but not implementation itself....
> Hi @potiuk @ashb ! Can we please check the changes again? Is there a way for someone else to review them ?:) I think it would really be sensible...
removed stale.
> @potiuk Why do we want to send a SIGHUP first? That seems like a very odd signal to send. SIGINT followed by SIGTERM is what Docker/podman/every other container runtime(?)...
> So my vote is just TERM -> ...timeout ... -> KILL ala Docker. The simpler, the better, yes. I am ok with it - as long as we also...
Also we should document (alongside the escalation of signals) whether we send signal to the process or to process group and whether we have separate process group. This makes a...