[lxc-users] cron zombies inside containers

[Brandon Rochon]

Any advice of tracking down the cause of various zombies on my LXC
containers?

I used to have zombie problems on my LXC hosts as well, but this was the
LXC API bug that was fixed in 1.0.6 (just installed from trusty-proposed).

However my zombie problems inside my Trusty containers persist.  One that's
fairly consistent is cron.  It almost looks like the old cron bug that I'm
pretty sure was fixed ages ago:
https://bugs.launchpad.net/ubuntu/+source/cron/+bug/137785

If I kill -9 the main cron proc the zombies clear.  Restarting the
container will also clear them (though it'll take awhile to stop). However,
they usually come right back.  For whatever reason if I restart the
container enough times, they stop returning completely though.

root at lxccontainer04:~# ps waux |grep Z |grep cron
root     11641  0.0  0.0      0     0 ?        Z    23:05   0:00 [cron]
<defunct>
root     11642  0.0  0.0      0     0 ?        Z    23:05   0:00 [cron]
<defunct>
root     11734  0.0  0.0      0     0 ?        Z    23:10   0:00 [cron]
<defunct>
root     11735  0.0  0.0      0     0 ?        Z    23:10   0:00 [cron]
<defunct>
root     11814  0.0  0.0      0     0 ?        Z    23:15   0:00 [cron]
<defunct>
root     11815  0.0  0.0      0     0 ?        Z    23:15   0:00 [cron]
<defunct>
root     11816  0.0  0.0      0     0 ?        Z    23:15   0:00 [cron]
<defunct>

BTW, these same cronjobs work just fine from the command line and on normal
(non-container) machines don't produce any zombies.

Any advice on tracking down the cause of this is much appreciated.

/Brandon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20141009/941de15f/attachment.html>