Status: New
Owner: ----
Labels: Type-Defect Priority-Medium
New issue 1778 by [email protected]: 100 Trying goes to wrong client
http://code.google.com/p/mobicents/issues/detail?id=1778
What steps will reproduce the problem?
1. We have a B2BUA application
2. We have several SIPp scripts generating different call scenarios. All
are running on the same host, each on a different port. All are using UDP
(in this case). All are generating 10 calls per second.
3. Occasionally (approx. 1 call in 2500) we are seeing a 100 Trying sent to
the wrong client.
What is the expected output? What do you see instead?
All messages in one session go to the same client
What version of the product are you using? On what operating system?
1.3.2 on Apache Tomcat, running on CentOS 5.4
Please provide any additional information below.
This has only been seen with "100 Trying", I have not seen evidence of any
other messages being mis-routed. We are not deliberately generating 100s,
we are leaving this to the stack.
Attached is a trace of a call where this happens.
192.168.58.21 is the SIPp host
192.168.58.70 is the B2BUA
In this call the INVITE is re-transmitted. The B2BUA then generates two 100
Tryings - The first to the correct destination (Port 5063), the second to
the wrong destination (Port 5061).
NB There is another client running on Port 5061, (it's not TLS, it's UDP).
I suspect that the stack has attempted to generate a 100 to the client on
5061, and somehow inserted the Call-ID, tags etc from the session to the
5063 client.
Attachments:
Wrong_100.pcap 5.8 KB
