every single phone call is being terminated at ran...
# telephony
a
every single phone call is being terminated at random points while the active conversation is happening, FreeSWITCH receives SIP 480 from LiveKit, one example call id is SCL_bfCyCPgfEa3y, @refined-appointment-81829 can you check?
If there are concurrent calls, all of them are terminated at the same time
r
Do you have a PCAP with the 480 in it?
a
r
Which packet in that capture has the 480
a
wait, I will send you another pcap in a min
r
ok
a
sorry, it was not 408, livekit is ending all of our calls by sending 200, here is a pcap from freeswitch side
r
200 to end call?
a
yes
r
Which packet number are you looking at in that PCAP
a
Packet #593 - This is where LiveKit sends the BYE message Key packets: - Packet #175 (160.5s): Call starts - INVITE from Twilio - Packet #187 (162.9s): LiveKit answers with 200 OK - Packet #593 (536.1s): LiveKit sends BYE (terminates the call after 6 minutes 13 seconds) - Packet #594: FreeSWITCH responds with 200 OK to the BYE
another one. Packet #754 - LiveKit sends the BYE message Key packets: - Packet #171 (161.5s): LiveKit answers call with 200 OK - Packet #754 (690.6s): LiveKit sends BYE (terminates after 8m 49s) - Packet #755: FreeSWITCH responds 200 OK to BYE
I made 10 concurrent calls for testing. 1. All calls terminated around 6 minutes (361-377 seconds) 2. All BYE messages came from LiveKit (packets #737, 743, 748, 753, 763, 768, 774, 780, 787, 804) 3. Very consistent pattern - first call in was first call out 4. All terminations were clean BYE messages with no error codes
r
I am looking at our PCAP for this call-id
SCL_Wx3m8YQifZyu
and I do not see our server sending a
BYE
.
Do you have multiple calls going to same room name?
a
No, they have different names, for the 10 concurrent calls room names are these: SCL_KbT96BeEzPxf - call_freeswitch_uz_fS8xEsrbkiQF SCL_YyCb8f2JYv4a - call_freeswitch_uz_jCMw9YheDhKS SCL_4zLu9KbB9HBr - call_freeswitch_uz_pbqeKhA5ZGfB SCL_Vtz5wub88SC9 - call_freeswitch_uz_wsrUxszYqiFH SCL_d7P9WwSkjYW3 - call_freeswitch_uz_MP2QQbJY6aro SCL_tJvEqruppX9c - call_freeswitch_uz_MBntZA6iJe2e SCL_LCP9894MSHfz - call_freeswitch_uz_tphmKmWmimSw SCL_2tb8yYPnj3Qw - call_freeswitch_uz_3SHpP9hr8i5g SCL_hxYhuLXFDVAf - call_freeswitch_uz_DMP22fUcDYfL SCL_s8f6hewuhwTC - call_freeswitch_uz_jCMw9YheDhKS
can you please have a look at sip-capture-10-concurrent.pcap?
Screenshot 2025-11-21 at 09.06.31.png
r
ok, I looked I am not sure what is going on there. I am getting a second opinion form our SIP subject matter experts.
👍 1
Does the same issue happen if you are not using egress?
a
I didn't try without egress, let me try
r
Thanks. I am not sure it will have an effect but I do see in the examples I look at egress appears to be having an issue just before room gets deleted. I am not sure if that is causal or just a correlation.
Looks like something in your systems is calling the delete room API that is causing the call to hangup
a
Ah, let me check
It was a bug on our side, thanks for checking