[Xymon] Xymon crashes

[Henrik Størner]

> So I uninstalled Xymon, deleted everyting related to it and then installed
> a fresh copy of Xymon 4.3.2 (also via the port-collection)
> The installation goes fine, but when I start it up, the html files aren't
> generated.
> It seems that xymond crashes for reasons I don't understand
>
> I did some digging and I see the following:
>
> xymonlaunch.log:
>
> 2011-04-09 23:39:33 xymonlaunch starting
> 2011-04-09 23:39:33 Loading tasklist configuration from
> /usr/local/www/xymon/server/etc/tasks.cfg
> 2011-04-09 23:39:33 Cannot open directory ^T<C2> (<C6><B1>

Looks like something weird in your configuration - that "directory" name 
doesn't look right. Check your configuration files - especially the 
xymonserver.cfg and tasks.cfg files, this is something that happens very 
early in the setup.

Maybe tasks.cfg has a "directory" include pointing to a non-existing 
directory. Maybe xymonserver.cfg is missing or unreadable.

Maybe your config files are not ascii, but in UTF-16 format.


> 2011-04-09 23:39:33 Setting up xymond channels
> 2011-04-09 23:39:33 Setting up logfiles
> 2011-04-09 23:39:38 Cannot open env file
> /usr/local/www/xymon/client/etc/xymonclient.cfg - No such file or
> directory
> 2011-04-09 23:39:39 Task xymond terminated by signal 6
> 2011-04-09 23:39:39 Loading hostnames

Are you logging all tasks to the same logfile ? It looks like it, and 
this makes it difficult to separate out which task logs an error.

xymond shouldn't be reading xymonclient.cfg - it should use 
xymonserver.cfg ...


> Output from gdb bin/xymond tmp/xymond.core
>
> #0 0x281afe17 in kill () from /lib/libc.so.7
> #1 0x281afd76 in raise () from /lib/libc.so.7
> #2 0x281ae98a in abort () from /lib/libc.so.7
> #3 0x0805df23 in sigsegv_handler (signum=11) at sig.c:57
> #4 <signal handler called>
> #5 rbtFind (h=0x0, key=0x2822b047) at rbtr.c:380
> #6 0x0805a461 in knownhost (hostname=0x2822b047 "Amp2hi-test",
> hostip=0xbfbfbcfc "", ghosthandling=GH_LOG) at loadhosts.c:344
> #7 0x0804d626 in get_hts (msg=Variable "msg" is not available.) at
> xymond.c:1051
> #8 0x08051be0 in do_message (msg=0x28204100, origin=0x806238f "") at
> xymond.c:2981
> #9 0x08056871 in main (argc=8, argv=Cannot access memory at address 0x4)
> at xymond.c:4999
>
> Line 6 surprises me, because 'Amp2hi-test' is one of the hosts that I used
> to monitor, but I removed all the old configs etc from the server.
> I've started up xymon several times and sometimes I see other 'old'
> hostnames in the output file of xymond.core
> It appears that I didn't completely remove all traces of the old
> installation after all...

Yep, something's lying around somewhere.



Regards,
Henrik