Random endpoint matching failure

Hello,

On a Debian-packaged asterisk 16.28.0 instance, I randomly loose incoming calls while outgoing calls are OK.

On a recent incoming call, console prints events like (redacted):

[Jan 22 10:29:03] NOTICE[2427]: res_pjsip/pjsip_distributor.c:676 log_failed_request: Request 'INVITE' from '"John Doe" <sip:123456789@192.168.77.54>' failed for '1.2.3.4:5060' (callid: f9815455-a970-4afc-9f88-1801b8e8452c) - No matching endpoint found

At the same time, pjsip show contacts or pjsip show registrations both show normal output.
When I issue ‘pjsip reload’, everything works OK again.

The PJSIP trunk with sometimes failing incoming calls, is (partly) defined by:

[main-trunk]
type=identify
endpoint=main-trunk
match=voip.itsp.com

At the moment of last matching failure, Debian name resolution seems OK.
I can’t say anything about the status of name resolution used by Asterisk to match voip.itsp.com with 1.2.3.4.

How can I debug this ?

Best regards

The hostname is resolved at load or reload time, if the hostname changed then it would not be reflected until such time as a reload occurred.

This trunk has a 120s registration period.
I suppose name resolution is updated during outbound registration.

I’ll dig deeper to gather more input.

Thanks for helping.

As I know resolution is only done on start/reload, no update during
registrations


Daniel