502 unavailable after executing pkill Passenger

I try to execute pkill Passenger to kill some background process tasks (watchdog, some ruby process, etc) and expect they will be recovered automatically. Turn out it was not, and now it just get stuck at nginx 502 error. What I suppose to do to make it right again?

We have a cronjob that will stop PUNs that are idle. I believe it’s every hour, though I’d have to check. Look at /etc/cron.d/ood for the job itself. You may increase it’s frequency to stop idle PUNs more aggressively.

That said, I think you’d need to stop the top level nginx processes. You should use /opt/ood/nginx_stage/sbin/nginx_stage helper to stop the PUNs. All the processes’ they start are required, so you have to stop the entire nginx stack, not just one or two processes.

I see, that explains why it back to normal automatically 1 hour later.

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.