bizzare application crash

[nmeyers at javalinux.net (Nathan Meyers)]

On Wed, Oct 09, 2002 at 02:22:21PM -0400, FRamsay at castelhq.com wrote:
> 
> I'm having a major problem with an application crashing.
> 
> The problem is that no core file is produced so I don't have a clue _why_
> it's happening.  I'm hoping someone has seen
> this kind of behaviour.
> 
> - there is no core file produced
> - the launch script explicity sets the coredump size to unlimited
> - the directory where the application is run from is writeable.
> 
> Since the above conditions are true, I'm thinking it's not actually
> producing a core file.  So my question is this, when would a crash _not_
> produce a core?

One possible answer: when it ends due to a signal that doesn't cause
a core dump - although those don't tend to look like what you'd call
a crash. The crash-related signals, like SIGFPE, SIGILL, and SIGSEGV,
trigger dumps.

Another possible answer: if it performs a chdir() to an unwritable
directory (or a directory you don't know to look in :-).

Nathan Meyers
nmeyers at javalinux.net


> 
> Kernel 2.4.9-31
> RedHat 7.2
> Linux Streams 2.13.16
> 
>                               -fjr
> 
> 
> 
> 
> Frank Ramsay
> Systems Programmer
> Castel, Inc
> 100 Cummings Center
> Suite 157h
> Beverly, MA 01915
> (978) 236 1000 (voice)
> (978) 236 1197 (fax)
> Email: framsay at castelhq.com
--