I wrote this message a few days ago:
http://openser.org/pipermail/devel/2006-December/005103.html
I am sorry for the lack of informations I gave, I could investigate a little bit about
it. Here is the full context:
When an INVITE comes, I dispatch and t_relay it to a server.
If the transaction makes an error, I wrote in the configuration to manage this whith a
failure route.
In this failure route, I dispatch and t_relay the transaction to a server for an other
service.
And then appears an strange error :
* with an openser compiled with the default flags, the UDP that received the last UDP
packet infinitely loops on the "sched_yield" function, in the "tsl"
function in fastlock.h.
* with an openser compiled with flags that make him use pthread for locking purposes, the
UDP process hangs, while trying to get a mutex.
It appears I forgot a call to append_branch in the failure route. Maybe the process tried
to get the same lock twice.
Is it known for OpenSER 1.1.0? If it is, is it corrected in later stable releases ?
Thanks a lot for your answer!
JF Smigielski.
________________________________________________________________________
iBELGIQUE, exprimez-vous !
http://web.ibelgique.com/