Help with zombie processes

[dictum9]

This is an Itanium box running 11.23. I see lots of zombie processes which are running. Is there a problem with the sendReport perl script? It works with sockets.


ps -fe | grep sendReport


root 24353 24343 0 Jun 19 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 24343 1 0 Jun 19 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 5556 1 0 May 29 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 16934 1 0 Jun 8 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 6756 6750 0 May 31 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 12644 1 0 Jun 6 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 4712 4704 0 May 30 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 25970 1 0 Jun 18 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 11792 1 0 May 27 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 7953 7949 0 May 24 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16494 1 0 May 23 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 4704 1 0 May 30 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 29876 29870 0 Jun 11 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 28185 28176 0 Jun 9 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 7949 1 0 May 24 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 12647 12644 0 Jun 6 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 29870 1 0 Jun 11 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 16401 16395 0 Jun 21 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 25979 25970 0 Jun 18 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16502 16494 0 May 23 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 27611 27604 0 May 27 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 18734 1 0 Jun 4 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 22601 22599 0 May 25 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16938 16934 0 Jun 8 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 6750 1 0 May 31 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 5559 5556 0 May 29 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16395 1 0 Jun 21 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 3751 3746 0 May 26 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 21678 21672 0 Jun 19 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 21672 1 0 Jun 19 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 27604 1 0 May 27 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 22599 1 0 May 25 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 18740 18734 0 Jun 4 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 3746 1 0 May 26 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 11795 11792 0 May 27 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 28176 1 0 Jun 9 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 26505 26500 0 Jun 12 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 26500 1 0 Jun 12 ? 0:00 sh -c /opt/legatoreport/bin/sendReport

[dictum9]

Another attempt at formating:

root 24353 24343 0 Jun 19 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 24343 1 0 Jun 19 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 5556 1 0 May 29 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 16934 1 0 Jun 8 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 6756 6750 0 May 31 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 12644 1 0 Jun 6 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 4712 4704 0 May 30 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 25970 1 0 Jun 18 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 11792 1 0 May 27 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 7953 7949 0 May 24 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16494 1 0 May 23 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 4704 1 0 May 30 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 29876 29870 0 Jun 11 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 28185 28176 0 Jun 9 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 7949 1 0 May 24 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 12647 12644 0 Jun 6 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 29870 1 0 Jun 11 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 16401 16395 0 Jun 21 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 25979 25970 0 Jun 18 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16502 16494 0 May 23 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 27611 27604 0 May 27 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 18734 1 0 Jun 4 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 22601 22599 0 May 25 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16938 16934 0 Jun 8 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 6750 1 0 May 31 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 5559 5556 0 May 29 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 16395 1 0 Jun 21 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 3751 3746 0 May 26 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 21678 21672 0 Jun 19 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 21672 1 0 Jun 19 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 27604 1 0 May 27 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 22599 1 0 May 25 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 18740 18734 0 Jun 4 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 3746 1 0 May 26 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 11795 11792 0 May 27 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 28176 1 0 Jun 9 ? 0:00 sh -c /opt/legatoreport/bin/sendReport
root 26505 26500 0 Jun 12 ? 0:00 /usr/bin/perl /opt/legatoreport/bin/sendReport
root 26500 1 0 Jun 12 ? 0:00 sh -c /opt/legatoreport/bin/sendReport

[Sandman! · Accepted solution]

Well they aren't zombie anymore since they are owned by init. But if init has been hanging onto them for some time then that could be a sign of a bigger problem. Might be best to open a support call with HP. Does a reboot fix this situation?

[Steven E. Protter]

Shalom,

Your output shows no indication of zombies.

How did you determine there were zombie processes? top?

If there are indeed zombies, nothing fixes them except a reboot. Normally its bad application code that creates zombies.

If you have a situation where zombies came back to life, you may have witnessed a computing miracle.

Please follow up with this. Inquiring itrc people want to know.

:-)

SEP

[Dennis Handly]

>Another attempt at formating:

You probably should attach a file.
Use the following to get hierarchical listing:
$ UNIX95= ps -Hef
(Adding -x will give you a longer command line.)

Are these reparented processes taking up any CPU time? If not, you might as well kill them. May and June is a long time ago.
It appears the perl script is hanging so sendReport waits.

>SEP: nothing fixes them except a reboot.

You kill zombies by killing the zombie master.
But you do have to read the fine print on the wanted poster to figure out if the master worth more dead than alive. :-)

[skt_skt]

Do you have a log file for the script running to identify where it is stucked.
Kill all of them and watch the log of next failure/hang