[hobbit] Re: Status not available

[Henrik Stoerner]

On Mon, Feb 28, 2005 at 03:13:05PM -0800, Brian Lynch wrote:
> Tried the compile with MEMORY_DEBUG and got this stacktrace from the core dump:
> 
> bash-2.03$ gdb bin/hobbitd core 
> GNU gdb 6.0
> Program terminated with signal 6, Aborted.
> #1  0xff13598c in abort () from /usr/lib/libc.so.1
> #2  0x000270b8 in xsprintf (dest=0x2e378 "xsprintf: Bogus
> destination\n", fmt=0x0) at memory.c:334
> #3  0x00013ad4 in posttochannel (channel=0x2e378, channelmarker=0x0,
> msg=0x0, sender=0x0,
>     hostname=0x3f2e8 "", log=0x0, readymsg=0x0) at hobbitd.c:436
> (gdb)

This looks really weird. 

Line 436 of hobbitd.c can only be reached if "readymsg" is not NULL.
But the call trace shows it as such.

The posttochannel() routine is always called with a message and a
sender. Yet both of those parameters show up as NULL in the call
trace.

I'd like to have a copy of this core-dump, the
~hobbit/server/bin/hobbitd binary, your entire /var/log/hobbit/
directory, the ~hobbit/server/tmp/hobbitd.chk file, and your
~hobbit/server/etc/bb-hosts and hobbitserver.cfg files.

Yes, this includes a bunch of info about your locally used ip-adresses
and hostnames. Check with your boss if it is ok to hand them out.


Is anyone else seeing repeated crashes of hobbitd ? I haven't heard of
any - quite the contrary - so I'm wondering why this happens so much
in just your installation.


Regards,
Henrik