our platform has been stable for quite some time.
we installed module updates and yum update’d on Saturday, then rebooted the server.
Since the server came back online, we are seeing an up-to 15 seconds delay in 2way RTP from our softphones / remote VPN users.
that is for an outbound call from a remote user:
SIP Signalling seems to flow properly
softphone sends the invite, 100 trying… 183 session progress
softphone SENDS rtp, phone server receives
trunk establishes 2way RTP
freepbx does not begin sending an RTP stream to the softphone.
Softphone does not receive p-early-media / trunk ringback.
–if we stay ringing for long enough, with no SIP signalling, the RTP starts flowing.
– if we answer the call, with no sip signalling, the rtp might start flowing
here are the oddities:
Softphone A, to Cellphone B
- sporadically works from home internets (4 different networks tested) over VPN
- always works while softphone is on the same subnet as the phone server
this affects both SIP and PJSIP users
good and failing users have Identical settings in Freepbx (via stare and compare)
What really confuses me is that comparing the SIP / SDP for a good call from A-B and a bad call from A-B, the syntax, rtp ports, IPs, etc. are identical. using the same phone, the same trunk, the same target.
Very confused, but would appreciate any attempts at helping troubleshoot this.