[plug] GDM hanging, leaving zombies

[Craig Ringer]

Hi folks

I'm still having trouble with my LTSP server's display manager. I 
thought it'd gone away, but clearly not - as running right now, GDM is 
not accepting or responding to XDMCP requests.

So - I'm going to ask for ideas, before I throw it out the window, stomp 
on it a few times, and start using XDM.

I'm seeing some defunct processes in the process list:
root     23707  0.0  0.0 16460 1168 ? S    08:44   0:00 [gdm-binary]
root     24111  0.0  0.0 16464 2004 ? S    08:58   0:16 [gdm-binary]
root     24250  0.0  0.0 16460 1172 ? S    09:05   0:00 [gdm-binary]
root     24415  0.0  0.0 16460 1172 ? S    09:16   0:00 [gdm-binary]
root     24730  0.0  0.0     0    0 ? Z    10:07   0:02 [gdm-binary 
<defunct>]
root     26835  0.0  0.0     0    0 ? Z    13:18   0:00 [gdm-binary 
<defunct>]
root     26999  0.0  0.0 16464 1196 ? S    13:27   0:01 [gdm-binary]
root     24242  0.0  0.0 13324 1160 ? T    Oct17   0:23 [gdm-binary]

and the primary GDM, pid 24242, is not spawining new children to handle 
new displays when it gets XDMCP requests. I've tried tracing pid 24242 
using strace -p, and strace simply hangs after attaching. Kill -9 is 
required to terminate it. gdb does the same thing. That's why the 'T' 
process status.

Because of this, I can't even get a backtrace or anything else to find 
out what's going on.

There are no unusual messages in the log. GDM's memory use seems 
reasonable at 1160k RSS, 13mb virtual including libs. There are just two 
zombie children, and a non-responding, non-traceable parent.

If anybody knows how to get more inforamation from processes behaving 
like this, I'd love to hear it. I'm going to have to kill -9 the lot of 
'em this afternoon after everybody goes, so that my users can get X 
sessions tomorrow.

The other issue is whether I can kill the parent and have the children 
survive. Ideally, I'd be able to do this - so that restarting the master 
gdm process didn't terminate all running XDMCP sessions - but it looks 
like they'll all die when the parent does :-(

I'm using gdm2 (2.4.0.7-14) on Red Hat 8. glibc version is 2.3.2-4.80.6, 
  gtk2 is 2.0.6-8, libgnome is 2.0.2-5. The server is a dual 2.4GHz Xeon 
with 2 gigs of RAM and 4 gigs of swap. Other info available if needed.

While I don't really /need/ GDM, it's actually very nice to work with 
except for this problem so I'd like to be able to fix it and keep going.

So.... *arrggh* Help!

Craig Ringer


_______________________________________________
plug mailing list
plug at plug.linux.org.au
http://mail.plug.linux.org.au/cgi-bin/mailman/listinfo/plug