Summary:ASTERISK-13740: Asterisk crashes when trying to leave a message via advanced mailbox options
Reporter:Ove Aursand (aurs)Labels:
Date Opened:2009-03-13 05:51:37Date Closed:2011-06-07 14:00:34
Versions:Frequency of
Environment:Attachments:( 0) backtrace.txt
Description:User dials in to his own mailbox, presses 3 for advanced options, then 5 to leave another user a message. When user enters mailbox id, asterisk sometimes crashes.

Asterisk, using realtime voicemail. Realtime engine is ODBC and database is postgresql.

Voicemail storage is via NFS, and asterisk is started with -t flag.

****** STEPS TO REPRODUCE ******

Dial in to a mailbox, press 3 for advanced options, then 5, and input a mailbox id to leave a message.
Comments:By: Tilghman Lesher (tilghman) 2009-03-13 11:46:11

This crash originates in the Postgres ODBC driver.  Please report this issue (along with backtrace) to that project.

By: Ove Aursand (aurs) 2009-03-13 18:42:33

Will test with the driver mentioned in http://bugs.digium.com/view.php?id=13300 and report results.

By: Ove Aursand (aurs) 2009-03-16 03:16:44

Unable to reproduce crash when using obdc driver from postgresql-odbc package in Centos. Similar issue also discussed here: http://mailman.unixodbc.org/pipermail/unixodbc-dev/2007-May/001036.html
It would be great if lmadsen could report to this bug what the conclution of that discussion was. Quote: "OK, after not having any crashes for the weekend utilizing the psqlodbc.so file in my /etc/odbcinst.ini file, I just had the same crash as previously it looks like."

By: Leif Madsen (lmadsen) 2009-03-16 07:41:00

I haven't run into that issue for quite some time, because the client I was using the PostgreSQL backend is no longer a client of mine. Since then, I have been developing using MySQL, so I don't even remember how I was producing that issue.

(looks at backtraces)

Ok, it looks like it had something to do with using sipfriends in the realtime, but other than that, I haven't run into the issue in quite some time. The system appears to have stabilized, but unfortunately I can't tell you exactly what was changed in order to resolve it. I don't believe anything was done in Asterisk -- probably just updating libraries and finding a combination that was stable.

By: Leif Madsen (lmadsen) 2009-03-16 07:56:33

Closing this issue after discussion with reporter on IRC. He will return if he can reproduce with the latest drivers, but as Tilghman states, this appears to be a crash in the driver itself, and not necessarily with Asterisk.