[asterisk-users] PJSIP Losing knowledge of external_media_address

[Mark Murawski]

I've seen this happen three times in the wild now.  I've been trying to 
isolate the source of the issue, but so far it seems like there's not 
enough debug output to know why this occurs.

Long story short:
- Start Asterisk
- PJSIP Handles receiving INVITE from ITSP via WAN (Asterisk is behind 
NAT).  SIP is handled correctly, Asterisk responds OK with RTP media 
address of external_media_address
- After 30 minutes to an hour or sometimes months later after startup, 
upon receiving INVITE from ITSP via WAN, Asterisk responds OK with 
INTERNAL LAN IP instead of external_media_address
- I've observed this occur after 30 minutes from startup with no 
configuration changes that were made or any pjsip reloads done during 
this period

pjsip
-------------
[global]
endpoint_identifier_order = username,ip,anonymous

[system]
type=system
threadpool_initial_size=30
threadpool_auto_increment=5
threadpool_idle_timeout=0
threadpool_max_size=100

[transport-udp]
type                       = transport
symmetric_transport        = yes
protocol                   = udp
bind                       = 0.0.0.0:5060
external_media_address     = 152.X.Y.Z
external_signaling_address = 152.X.Y.Z
external_signaling_port    = 5060
allow_reload               = no
tos                        = cs3
cos                        = 3
local_net                  = 127.0.0.1/24
local_net                  = 192.168.50.0/24
local_net                  = 192.168.1.0/24
local_net                  = 10.3.2.0/24
local_net                  = 10.20.1.0/24
local_net                  = 10.10.41.0/24
local_net                  = 10.5.1.0/24

pjsip_wizard
-------------

[isoft-sr-in-1]
type = wizard
transport = transport-udp
remote_hosts = 192.81.237.20
aor/max_contacts = 1
aor/remove_existing = yes
aor/qualify_frequency = 60
aor/qualify_timeout = 2000
endpoint/ice_support = no
endpoint/disallow = g723,slin,ilbc,lpc10,g729,speex,g726aal2,g722
endpoint/allow = ulaw,alaw,adpcm,gsm
endpoint/direct_media = no
endpoint/force_rport = yes
endpoint/rewrite_contact = yes
endpoint/rtp_keepalive = 30
endpoint/rtp_symmetric = yes
endpoint/rtp_timeout = 60
endpoint/rtp_timeout_hold = 60
endpoint/send_pai = yes
endpoint/send_rpid = yes
endpoint/trust_id_inbound = yes
endpoint/trust_id_outbound = yes
endpoint/trust_connected_line = no
endpoint/send_connected_line = no
endpoint/context = trunkhandler_pbx-sip-t1


Attached sip sessions and debug log... the only thing I found 
interesting was finding a lack of a log item
We SHOULD be seeing:
DEBUG[XXXXX] res_pjsip_session.c: (null session): Setting external media 
address to 152.X.Y.Z
This message is clearly lacking from the debug session where the 
incorrect media address is sent.  But there's not enough detail in the 
debugs to see why this decision was not made to use external_media_address

By default we use nat settings for all our endpoints, but obviously it's 
not required here for an ITSP that has trustworthy media ports in the 
SDP.  Maybe a bandaid is turning off rewrite_contact for this endpoint?  
Going to try that as soon as possible.

Why is external_media_address not being used all of a sudden?  Has 
anyone else seen this... is this a bug?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: debug.log
Type: text/x-log
Size: 28770 bytes
Desc: not available
URL: <http://lists.digium.com/pipermail/asterisk-users/attachments/20230818/b65ab0fa/attachment-0001.bin>