Closed kadel closed 5 years ago
I can see this also with nodejs
16 1716 ? Ss 17:17 0:00 /bin/sh
1000160+ 18513 0.0 0.0 51704 1692 ? R+ 17:17 0:00 \_ ps afxu
1000160+ 1 0.4 0.2 360060 9244 ? Ssl 17:08 0:02 /var/lib/supervisord/bin/supervisord -c /var/lib/supervisord/conf/supervisor.conf
1000160+ 18275 0.8 0.0 0 0 ? Z 17:17 0:00 [npm] <defunct>
1000160+ 18299 0.4 0.0 0 0 ? Z 17:17 0:00 [node] <defunct>
1000160+ 18378 0.6 0.0 0 0 ? Z 17:17 0:00 [npm] <defunct>
1000160+ 18398 0.3 0.0 0 0 ? Z 17:17 0:00 [node] <defunct>
1000160+ 18471 0.0 0.0 11688 1408 ? S 17:17 0:00 /bin/bash /var/lib/supervisord/bin/setup-and-run
1000160+ 18480 1.9 0.7 733196 28476 ? Sl 17:17 0:00 \_ npm
1000160+ 18500 1.4 0.6 568648 25772 ? Sl 17:17 0:00 \_ node app.js
might be related to https://github.com/ochinchina/supervisord/issues/60
I've quick fix: https://github.com/redhat-developer/odo-supervisord-image/tree/fix-zombie-problem https://github.com/kadel/odo/tree/fix-zombie-problem
The problem is that we run supevisorD as pid1, pid1 should be also responsible for cleaning ophraned process, but supervisorD doesn't do that. Running https://github.com/Yelp/dumb-init as pid1 solves that problem.
restarting httpd via supervisor leaves zombie processes behind