daemon.py is creating a zombie process

Bug #206960 reported by h3xis
24
This bug affects 4 people
Affects Status Importance Assigned to Milestone
wicd
Confirmed
Low
Unassigned

Bug Description

I have Wicd 1.4.2 installed on Debian Lenny from the repo. I'm not sure what's going on but daemon.py is creating a zombie process called 'sh'.

Revision history for this message
Dan O'Reilly (oreilldf) wrote :

I noticed this with 1.4.2 and lower as well. I haven't experienced it while I've been running the 1.5.0 daemon (though I haven't been paying close attention to it), so there's a good chance it's fixed. Not 100% sure though, once 1.5.0 gets released give it a shot and see how it runs.

Dan O'Reilly (oreilldf)
Changed in wicd:
importance: Undecided → Low
status: New → Confirmed
Revision history for this message
Dan O'Reilly (oreilldf) wrote :

Now that we're not using a shell to spawn off sub-processes this shouldn't be an issue.

Changed in wicd:
status: Confirmed → Fix Released
Revision history for this message
cstorm (christian-storm) wrote :

I'm using wicd 1.6.2 and dhcpcd 5.0.4 on archlinux and get a dhcpcd zombie:
python -O /usr/lib/wicd/wicd-daemon.py
   /usr/bin/python -O /usr/lib/wicd/monitor.py
   [dhcpcd] <defunct>
wicd is running in daemon mode. There's no hint in the log file or syslog about it.
There's also an archlinux forum post about this issue: http://bbs.archlinux.org/viewtopic.php?id=74952

Revision history for this message
h3xis (h3xis) wrote :

I can confirm this in 1.6.1.

Changed in wicd:
status: Fix Released → New
Revision history for this message
Andrew Psaltis (nacl) wrote :

Observed in 1.6.2.

Changed in wicd:
status: New → Confirmed
Revision history for this message
Robby Workman (rworkman) wrote :

I'm seeing a defunct dhcpcd process here as well. I thought it might have been something peculiar about dhcpcd-3.2.x, but I'm also seeing it with dhcpcd-5.0.6. I don't even know where to begin looking for the problem -- I don't see any obvious issue with how wicd calls dhcpcd...

Revision history for this message
Andrew Psaltis (nacl) wrote :

I've just observed this in 1.6.2 with dhclient. It seems that something is keeping the original thread alive despite the fact that it should have forked into the background (which is what the live thread is) and exited normally.

Revision history for this message
Paul Loughman (snowhog) wrote :

Kubuntu Karmic 9.10, kernel 2.6.31-20-generic #57-Ubuntu SMP Mon Feb 8 09:05:19 UTC 2010 i686 GNU/Linux, KDE 4.4.1, wicd 1.6.2.2-1, dhcpcd 1:3.2.3-3ubuntu1 (not installed).

top shows:

top - 13:29:29 up 4:44, 2 users, load average: 0.09, 0.04, 0.06
Tasks: 191 total, 1 running, 188 sleeping, 1 stopped, 1 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 2052252k total, 1951088k used, 101164k free, 129696k buffers
Swap: 2096440k total, 0k used, 2096440k free, 1324188k cached

sudo ps -elf | grep Z shows:

F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD
4 Z root 2070 1643 0 80 0 - 0 exit 08:45 ? 00:00:00 [dhclient] <defunct>
1 S paul 23982 21197 0 80 0 - 17605 poll_s 13:21 ? 00:00:00 kdeinit4: kio_http [kdeinit] http local:/tmp/ksocket-paul/klauncherT21198.slave-socket local:/tmp/ksocket-paul/konquerorZ21478.slave-socket
0 S paul 24180 23998 0 80 0 - 761 pipe_w 13:30 pts/1 00:00:00 grep --color=auto Z

The zombie is always [dhclient] <defunct> and although it isn't hurting anything, and easy to kill (sudo kill -9 [ppid #]), it shouldn't be created.

Revision history for this message
Sheldon Hearn (sheldonh) wrote :

Happens with 1.7.0 too.

Revision history for this message
Michael Kory Woods (kory-i) wrote :

I can confirm it's a "problem" with Arch running wicd 1.7.0 and dhcpcd 5.2.12; both restarting the wicd daemon and "kill --9"'ing takes care of it. It happens after the computer returns from standby (using acpid to suspend to memory).

Could swear it also happened when the computer finished booting, as well; however, I'll have to confirm that for certainty.

Revision history for this message
Michael Kory Woods (kory-i) wrote :

It's also a problem when initially starting (hence, without having to suspend/resume).

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.