On Wed, 2010-10-06 at 12:28 -0500, Vernon A. Fort wrote: > On Wed, 2010-10-06 at 13:22 -0400, Victor Duchovni wrote: > > On Wed, Oct 06, 2010 at 12:15:36PM -0500, Vernon A. Fort wrote: > > > > > On Mon, 2010-10-04 at 17:30 -0400, Victor Duchovni wrote: > > > > On Mon, Oct 04, 2010 at 03:43:44PM -0500, Vernon A. Fort wrote: > > > > > > > > > I've noticed this: > > > > > > > > > > mail.log.1:Oct 3 03:32:43 ns postfix/master[3672]: warning: > > > > > process //usr/lib/postfix/postscreen pid 19122 killed by signal 11 > > > > > > > > > > and > > > > > > > > > > [457795.641083] postscreen[19122]: segfault at 1f ip 0804c463 sp > > > > > bfcf5e50 error 4 in postscreen[8048000+29000] > > > > > > > > Enable "core" file generation and report a backtrace. > > > > > > > > > > I have a core and backtrace file. I tar.gz but i keep getting message > > > 4000 chars. how do i get this info to the list? > > > > We don't yet need the whole core file, just run: > > > > gdb /usr/sbin/postscreen /path/to/core > > ... copious output ... > > gdb) bt > > > > The "bt" command generates a back-trace (stack trace), that should show > > where the segfault occurred. Ideally, your postscreen is compiled with > > "-g" or "-g -O", i.e. with symbols and line numbers, ... > > > > attached. This is a gentoo install so added the "ggdb" flag to the > make.conf file and re-compiled postfix. I also removed the > -fomit-frame-pointer per the backtrace doc's. > > This is my first time working with core and backtrace files - bear with > me. > > Vernon
I'm thinking this is no good - do we need debugging symbols? this is from gdb /usr/lib/postfix/postscreen "corefile". Reading symbols from /usr/lib/postfix/postscreen...(no debugging symbols found)...done. Vernon
