Hi,
o Jeremy A [08/14/09 02:53]:
Hi,
I have an application running in sems that works perfectly when using
unixsockctrl interface but when I use sipctrl interface I get partial
SIP packet loss (and/or ?) delay when the traffic rate is higher than
normal.
The sipctrl interface works well if I don't put a lot of SIP traffic
through it, but partially fails when the load increases. The rate where
failure occurs is a few tens of packets in a single burst. The problem
triggers when I do a simultaneous dialout to 5-10 different targets
using a separate dialog class object/thread for each dialout. The
failure pattern is not repeatable between tests - i.e. sometimes it
works sometimes it doesn't
Before I try and find the problem, is this a known issue? If so, is
not anything that I am aware of. Do you get any error messages in the
logs? Do you see retransmissions of the INVITEs sent by sipctrl in those
cases?
Apart from taking a little longer until all INVITEs are sent out, all
messages should be sent, even if they are sent in a burst. If not, and
something fails, then it should at least return some error somewhere.
I rather had quite some trouble with unixsockctrl under load, thats the
reason we have been using almost only sipctrl for quite a while now, for
high and low load situations.
Stefan
there a fix in CVS head?
Thanks
Jeremy
_______________________________________________
Sems mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/sems
--
Stefan Sayer
VoIP Services
[email protected]
www.iptego.com
IPTEGO GmbH
Wittenbergplatz 1
10789 Berlin
Germany
Amtsgericht Charlottenburg, HRB 101010
Geschaeftsfuehrer: Alexander Hoffmann
_______________________________________________
Sems mailing list
[email protected]
http://lists.iptel.org/mailman/listinfo/sems
