Subject: | segfault on mysql_ping |
Hi,
I would like to report information about this problem (see also
http://bugs.mysql.com/bug.php?id=36810).
Using RT on Debian lenny, and fastcgi, we encouter sometimes this
problem (the fastcgi process segfault).
It's in the DBD::mysql ping function, here are the steps I used to
reproduce it:
- set low timeout in my.cnf (10s)
- start the app (RT)
* I show a first connect and ping
- wait for timeout using SHOW FULL PROCESSLIST
- connect to the app
* works fine, RT reconnects, ping works, adding debug in
mysql_db_reconnect function I see that "imp_dbh->auto_reconnect" is
false and so there is no automatic reconnect from DBD, but RT has code
to do it and does it.
- wait timeout another time
- reconnect to the app
* mysql_db_reconnect detect now that auto_reconnect is on (why???) try
to reconnect, return TRUE, then the next call to mysql_ping in the
DBD::mysql ping function segfault.
Please don't hesitate to ask if you need more debugging to solve this :)