Good day.
I think I have a lead on this to help come up with a resolution. I might be able to do it myself but I need to really dig into the code and kamailio to understand it better.
First, I was using the sip_trace module along with Homer to gather the info, but it did seem incomplete. The work of the SCA module is not picked up by the sip_trace module. I believe because it if doing instead script work and sending its own packets, which would not be picked up by the script anymore. I finally used sngrep (referred by Javan Strika who is also having the issue) and was able to see the entire conversations for the SCA's. But the sngrep in a "HEP" and "non verbose" mode seems to have a memory leak and will crash kamailio due to no more memory. So, I am now turning to captagent to get a more stable system to get better traces for this issue.

Now for the SCA problem. It seems that the unreleased "seized event" resides with the "held" event and the reinvite/invite that occurs to kamailio. It seems to get lost, hungup. pushed around or jacked up in the 'route[WITHINDLG]' section of the script and the reinvites/invites will keep happening until a timeout occurs or the sometimes I have seen the 404 Not here error occur.

I know this is not much to go on until I upload a good trace but maybe a brainiac might go...Oh, I know that problem and quicker fix.

Using Kamailio 4.2.7 with latest patches. CentOS 6.7-latest. One test bed is a mix of 2 Polycom VVX [34]10 with 5.4.4 and 1 Polycom 650 with 3.3.5; Another is a mix of 14 Polycoms VVX [34]10 with 5.3.2, another is a mix of 4 Polycom VVX [34]10 with 5.3.2

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.