topic cron issue in Operating System - HP-UX

cron issue

lawrenzo_1 — Mon, 11 Feb 2008 16:44:49 GMT

Hi all,

I am investigating an issue where several defunc progs are being produced between 1600 - 2200 on evry hour and half hour:

output from my last failure:

imx018> # ps -ef |grep 17417
root 13232 17417 0.0 - ?? 0:00.00
root 14125 17417 0.0 - ?? 0:00.00
root 14375 17417 0.0 - ?? 0:00.00
root 15199 17417 0.0 - ?? 0:00.00
root 15294 17417 0.0 - ?? 0:00.00
root 15953 17417 0.0 - ?? 0:00.00
root 16369 17417 0.0 - ?? 0:00.00
root 16633 17417 0.0 - ?? 0:00.00
root 16713 17417 0.0 - ?? 0:00.00
root 17236 17417 0.0 - ?? 0:00.00
root 17417 789 0.0 17:30:00 ?? 0:00.04 /usr/sbin/cron
root 17693 17417 0.0 - ?? 0:00.00
root 17804 17417 0.0 - ?? 0:00.00
root 18679 17417 0.0 - ?? 0:00.00

imx018> # ps -ef |grep cron
root 789 1 0.0 Jan 21 ?? 1:43.60 /usr/sbin/cron
root 17417 789 0.0 17:30:00 ?? 0:00.04 /usr/sbin/cron

as you can see 789 is the original cron process however when the defunc procs are produced another cron daemon is left hanging around.

Investigations I have done so far are:

checked all cron jobs that start between 00, 30 from the /var/spool/cron/crontabs directory.

- commented out all jobs that start at this time and the defunc still appear

- added two lines to each cron script scheduled to obtain the PID and COMMAND and sent this to a file, no pids match any of the defunc pids.

all I can think is maybe there is a job that is started by the original program that initiates another script which is failing however as I mentioned I have commented out all jobs ( one at a time ) and the defun procs still come .....

The final thing i performed was to restart the cron daemon.

Does anyone have any idea how I can trouble shoot this further?

there is no indication of the issue in mail or syslog.

thanks for any help.

Chris.

Re: cron issue

Dennis Handly — Mon, 11 Feb 2008 17:12:16 GMT

This shouldn't be happening. cron shouldn't be creating zombies (zombie master). If it is, then it is a bug in cron. Have you looked for any cron patches?

root 13232 17417 0.0 - ?? 0:00.00
I assume something like "ps -fp 13232" doesn't give you the name of the process?
It is root, does that tell you anything?

>The final thing i performed was to restart the cron daemon.

Did the restart create another zombie master?
Have you tried killing the zombie master 17417?

Re: cron issue

lawrenzo_1 — Mon, 11 Feb 2008 17:25:06 GMT

thanks Dennis,

I have checked the ps -fp and nothing exept the defunc proc being displayed.

running ./cron stop from init.d directory didnt kill the second cron master so killed manually

I restarted cron and have also removed a cron file for a user that had the home area removed / not present and didnt see any errors appear at 1700

offending user:

riviera:Nologin:xxx:xxx:Riviera transfer:/usr/users/tx/riviera:/bin/ksh

the directory /usr/users/tx/riviera is not present and although the cron jobs associated with this were not relevant - hopefully this will fix my issue ...

Thanks for the advise

Re: cron issue

lawrenzo_1 — Tue, 12 Feb 2008 10:25:55 GMT

identifying the reviera user and removing the cron file has resolved my issue.