master segfaults on Solaris 9
igor at ipass.net
Wed Feb 4 12:01:30 EST 2004
On Wed, 4 Feb 2004, Markus Wernig wrote:
> On Wed, 2004-02-04 at 15:42, Gary Mills wrote:
> > On Wed, Feb 04, 2004 at 09:15:13AM -0500, Igor Brezac wrote:
> > >
> > > On Wed, 4 Feb 2004, Markus Wernig wrote:
> > > >
> > > > I have uploaded a more complete truss file (truss -u *:*:*
> > > > /usr/cyrus/bin/master) to
> > > > http://markus.wernig.net/en/cyrstart.trussall.txt, which shows all
> > > > library calls as well as the system calls.
> > >
> > > A backtrace would be much more helpful. I use Solaris 9 without any such
> > > issues.
> > You can use `coreadm' to permit global core dumps and global setid
> > core dumps. I put them in /var/core, with the pattern suggested in
> > the man page. Then, you will get a core file when master exits.
> > `pstack' on the core would be interesting.
> Hi Gary
> Here you go (sorry, I'm not familiar with coredumps):
> # pstack master-5002-8822-1075912530-core
> core 'master-5002-8822-1075912530-core' of 8822:
> feec45a8 set_zone_context (40211f52, 78474, 0, 0, 0, 0) + 8c
> feec3ba8 localtime_r (ffbff27c, ffbfed9c, 0, 0, 0, 0) + 20
> feeb996c ctime_r (ffbff27c, ffbff260, 1a, 0, 0, 0) + c
> feed583c vsyslog (fef428d0, 0, ffbffa3c, 400, 0, 36504) + 1d4
> feed5658 syslog (5, 19728, 0, 40211f54, 39fc8, 0) + 1c
> 000156b8 main (1, ffbffbe4, ffbffbec, 36c84, 0, 2c400) + 488
> 00012238 _start (0, 0, 0, 0, 0, 0) + 5c
A gdb backtrace (with debugging compiled in) would be more helpful. This
does not look like the cyrus problem. Try to recompile cyrus.
More information about the Info-cyrus