anacron/run-parts failing with defunct processes
Recently, anacron has been hanging on one of my systems. It first
happened over a month ago and I didn't notice until 'locate' warned me
the database was over 7 days old and I saw 'run-parts /etc/cron.daily'
was hung. I killed anacron and ran reran by hand with no problems.
The next day, anacron hung again. That night I needed to power the
machine off (for a planned power outage at the office) and after the
machine was brought back up it stopped having the problem... until now.
On Friday, cron.daily hung again and has been hanging every day since.
For example, I manually ran 'anacron -n -d -f'. It seemed to take a
long time on dlocate, so I found the anacron PID (21111) and did:
# pstree -p|grep 21111
| |
-sshd(20942)---bash(20947)---anacron(21111)---run-parts(21112)---dlocate(21182)
I then ran 'ps axu|grep 21182' and saw:
root 21182 0.0 0.0 0 0 pts/11 ZN 07:30 0:00 [dlocate
<defunct>]
I then tried the following:
# cd /etc/cron.daily
# run-parts --verbose --report .
run-parts: executing ./0anacron
run-parts: executing ./acct
It didn't appear to be doing anything, so from another window, I ran:
# ps axu|grep acct
root 22429 0.0 0.0 0 0 pts/11 Z 08:13 0:00 [acct
<defunct>]
Running manually has hung at exim, dlocate, acct and htdig. At other
times, At other times, I've been able to execute 'run-parts' get
through everything, but usually it dies somewhere along to way.
The system is a dual Pentium II machine running Debian Woody with a
2.4.18 kernel, 255M RAM.
Reply to: