[Freeswitch-users] Revision 16238, 16237 compiling and call handling issue

[Ahmed Naji]

People,

It seems there are  a few things that are broken in this release. So far,
I've come across two issues: first of, configure fails, per what others
reported, complaining about NONE/bin not being an absolute directory path;
1637 has no problems with that.

There is also what appears to be a serious call handling issue that was not
present in 1.0.4 trunks which is the following:

Calls are initiated correctly, Leg-B is set-up and remote end rings,
ring-back is reported to Leg-A, remote end picks up, but Leg-A still gets
the ringing tone. This is pretty much the case with both internal and
external profiles, irrespective of whether or not a gateway is used to route
the call, irrespective of which media mode FS is running in and irrespective
of codec.

I know dev and support groups don't like to get involved in performance and
load testing loops, but for he who cares, performance on 1.0.5 trunks is
miserable compared to 1.0.4. I've bombarded 1.0.4 with over 150 cps with
hardly a glitch on quad-xeon machines with 6Gb RAM and GigE controllers.
1.0.5 falls far, far short of that and the figure is a fraction at just
under 20 for the same spec. machine. I run CentOS 5.0.4, and have had the
exact same experience with Debian stable and Ubuntu Karmic on the same
hardware.

I'm tracing the call handling issue now, and will report back if I find
anything useful. Meantime, it would be great if someone from support can
open a Jira for this or let me know how to do that - sorry, my own ignorance
of how the dev/support system works .. only converted to FS 6 weeks ago :).
Happy to post traces and findings.

Regards,

Ahmed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20100111/1764b469/attachment-0002.html