[HTCondor-users] condor_shadow producing zombie (defunct) processes (Condor 8.0.3)

[Antonio Chay <achay@xxxxxxxxxxx>]

We've detected some zombie processes produced by condor_shadow
Currently using:

CentOS release 6.4
condor-8.0.3-174914.x86_64 (from wisc.edu)

Zombies:
XXXX     10879  5889  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10884  5218  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10889  5517  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10893 32587  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10895  6151  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10902 32635  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10904  4381  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10910  6273  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10934  6217  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10939  5788  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10949  5210  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10964 32643  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     10994  4405  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     11005  5332  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     11031  6245  0 15:12 ?        00:00:00 [false] <defunct>
XXXX     11033  8530  0 15:12 ?        00:00:00 [false] <defunct>

Checking parents:
ps -p 5218 -f
UID        PID  PPID  C STIME TTY          TIME CMD
XXXX      5218 19870  0 15:06 ?        00:00:00 condor_shadow -f
958534.0 --schedd=<10.255.6.101:43216>

Init takes cares of them smoothly, but it looks suspicious.

This is the first time we've seen this in our cluster.

Should we be worried about it?
Should we upgrade to 8.0.4? (we've just pushed this new version on
October 14th, this is production environment).

Regards!
Antonio.