[OPEN-ILS-DEV] OpenSRF-C segmentation fault

[Dan Scott]

Hi James - welcome!

Just one thought here:

Have you tried "osrf_ctl ... -a restart_all" after the initial
segfault? There's currently a race condition at start up where, if the
Perl processes are not up and running by the time OpenSRF-C attempts
to connect opensrf.settings, OpenSRF-C will crash. Restarting all of
the servies should enable a succesful launch of OpenSRF-C the second
time because all of the required Perl modules will already be cached
in memory.

I see the race condition on VMWare, and wouldn't be surprised if your
AMD 350 is suffering from the same issue. I've got an idea about how
to prevent the condition from occurring, but just have to find enough
time to implement it properly...

Anyways, I hope this helps. It's a complex stack to get working
together at the moment, so troubleshooting is complex too :)

On 27/04/07, James Fournie <james.fournie at gmail.com> wrote:
> Hi all,
>
> I am attempting to do a test install of Evergreen on a clean
> "text-only" install from the Ubuntu-Alternate CD.  My machine (nesika)
> is a lowly AMD K7 350, which is only on an intranet.
>
> OpenSRF-C crashes with a segmentation fault and the log says:
>
> opensrf 2007-04-27 15:16:36 [INFO:4820:osrf_system.c:195:]
> Bootstrapping system with domain nesika, port 5222, and unixpath
> (null)
> opensrf 2007-04-27 15:16:36 [INFO:4820:osrf_app_session.c:515:]
> [opensrf.settings] sent 214 bytes of data to
> router at nesika/opensrf.settings
> opensrf 2007-04-27 15:16:36 [INFO:4820:transport_session.c:406:]
> Received <error> message with type cancel and code 503
> opensrf 2007-04-27 15:16:36 [WARN:4820:osrf_stack.c:84:]  !!! Received
> Jabber layer error message
> opensrf 2007-04-27 15:16:36 [WARN:4820:osrf_stack.c:94:]  * Jabber
> Error is for top level remote id [router at nesika/opensrf.settings], no
> one to send my message too!!!
> opensrf 2007-04-27 15:16:36 [INFO:4820:osrf_stack.c:103:] Message
> processing duration 0.000798
> opensrf 2007-04-27 15:17:36 [INFO:4820:osrf_app_session.c:142:]
> Returning NULL from app_request_recv after timeout
>
> I can telnet to the jabber server.
>
> TIA,
>
> James Fournie
> Digitization Librarian
> Union of BC Indian Chiefs
>
> PS: I had the following problems following the Ubuntu 7.04 Prereq
> installation guide, which may or may not be related to my problem
> above:
> #10 - I couldn't install DateTime::Format::ISO8601 from CPAN, so I
> used dh-make-perl/dpkg
> #15 - aclocal must be run from the libdbi dir
> #17 - aclocal must be run from the libdbi-drivers dir
> #19 - "host all 127.0.1.1/32 md5" doesn't work unless you add a second
> 'all' in there
> #20 - second bullet should be version 8.2 not 8.1
> #24 - should there be a semicolon after JSERVER=""  ?
>


-- 
Dan Scott
Laurentian University