Thread: Segfault in pgsql, Sparc Solaris 2.7, Postgresql 7.1.1
Howdy, I'm getting a segfault from psql when a createdb is attempted. Postgresql is configured with: ./configure --prefix=/opt/pgsql --enable-syslog and seems to compile without trouble (gcc 2.95.2), but fails during the createdb stage of the regressions tests. Installing and attempting to do the relevant bit of createdb from the command line: ====== $ gdb /opt/pgsql/bin/psql GNU gdb 4.18 <snip banner> (gdb) set args -d template1 -c "CREATE DATABASE \"paulmtest\"" (gdb) run Starting program: /opt/pgsql/bin/psql -d template1 -c "CREATE DATABASE \"paulmtest\"" CREATE DATABASE Program received signal SIGSEGV, Segmentation fault. 0xff363b60 in __do_global_dtors_aux () from /opt/pgsql/lib/libpq.so.2 (gdb) bt #0 0xff363b60 in __do_global_dtors_aux () from /opt/pgsql/lib/libpq.so.2 #1 0xff3728d4 in _fini () from /opt/pgsql/lib/libpq.so.2 #2 0xff3ba060 in ?? () #3 0xff0a0130 in _exithandle () from /usr/lib/libc.so.1 #4 0xff116964 in exit () from /usr/lib/libc.so.1 ====== Any ideas? -- Paul McGarry mailto:paulm@opentec.com.au Systems Integrator http://www.opentec.com.au Opentec Pty Ltd http://www.iebusiness.com.au 6 Lyon Park Road Phone: (02) 9870 4718 North Ryde NSW 2113 Fax: (02) 9878 1755 -------------------------------------------------------------------- This document and any attachments are intended solely for the named addressee(s), are confidential, and may be subject to legal professional privilege. Please notify us (on +61-2 9878 1744) as soon as possible if you have received this document in error. Any confidentiality or privilege is not waived or lost because this email has been sent to you by mistake. This document and any attachments are subject to copyright. No part of them should be reproduced or distributed by any means whatsoever without the prior consent of the copyright owner. Opentec does not warrant that this email and any attachments are error or virus free. --------------------------------------------------------------------
Hi Justin, > I'm not sure if it'll be of assistance with this, but there's > a Solaris > specific installation guide for PostgreSQL at : > http://techdocs.postgresql.org/installguides.php#solaris I've compiled postgres 6.5 and 7.0 so many times I just jumped in and went with it :) Upgrading to gcc 2.95.3 has fixed the problem I was having, it's compiled and the regression tests are now whirring away happily. Those build instrucions look good, as does the init script. Thanks again. -- Paul McGarry mailto:paulm@opentec.com.au Systems Integrator http://www.opentec.com.au Opentec Pty Ltd http://www.iebusiness.com.au 6 Lyon Park Road Phone: (02) 9870 4718 North Ryde NSW 2113 Fax: (02) 9878 1755 -------------------------------------------------------------------- This document and any attachments are intended solely for the named addressee(s), are confidential, and may be subject to legal professional privilege. Please notify us (on +61-2 9878 1744) as soon as possible if you have received this document in error. Any confidentiality or privilege is not waived or lost because this email has been sent to you by mistake. This document and any attachments are subject to copyright. No part of them should be reproduced or distributed by any means whatsoever without the prior consent of the copyright owner. Opentec does not warrant that this email and any attachments are error or virus free. --------------------------------------------------------------------
Paul McGarry <PaulM@opentec.com.au> writes:
> Program received signal SIGSEGV, Segmentation fault.
> 0xff363b60 in __do_global_dtors_aux () from /opt/pgsql/lib/libpq.so.2
> (gdb) bt
> #0 0xff363b60 in __do_global_dtors_aux () from /opt/pgsql/lib/libpq.so.2
> #1 0xff3728d4 in _fini () from /opt/pgsql/lib/libpq.so.2
> #2 0xff3ba060 in ?? ()
> #3 0xff0a0130 in _exithandle () from /usr/lib/libc.so.1
> #4 0xff116964 in exit () from /usr/lib/libc.so.1
This appears to be psql dying, not the backend. Renaud Thonnart
<thonnart@amwdb.u-strasbg.fr> reported a similar problem on his Solaris
setup last week (4-May in pgsql-general), but so far as I heard no one
had a clue what the problem was. <WAG> Shared library linkage problem,
maybe? </WAG>
regards, tom lane