Pjsip re-register creating downtime

Sangoma Hardware Phones
1

We’ve got P320 and P330 phones that are periodically choosing to delete and recreate their contact rather than re-up the timeout. It happens periodically every 45 minutes and some endpoints stay unavailable for minutes, others only a second. I’m seeking any advice on how to avoid this downtime, to either keep the phones refreshing their registration rather than re-creating it, or figure out what causes the downtime to be long in some phones but not others.

Typically these lines are preceeded by a lot of DPMA activity, presumably endpoint manager sending data using pjsip. When I looked into it with pjsip logger on, these messages were initiated by the endpoint.

Summary

[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:1] Set(“Message/ast_msg_queue”, “MESSAGE(custom_data)=mark_all_outbound”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:2] Set(“Message/ast_msg_queue”, “TMP_RESPONSE_URI=pjsip:172.18.17.36:5060;transport=udp”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:3] Set(“Message/ast_msg_queue”, “MESSAGE_DATA(Request-URI)=”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:4] Set(“Message/ast_msg_queue”, “MESSAGE_DATA(X-Digium-AppServer-Response-URI)=”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:5] Set(“Message/ast_msg_queue”, “MESSAGE_DATA(X-Digium-AppServer-Response-FullContact)=”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:6] MessageSend(“Message/ast_msg_queue”, “pjsip:172.18.17.36:5060;transport=udp,proxy”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Executing [digium_phone_module@dpma_pjsip_message_context:7] Hangup(“Message/ast_msg_queue”, “”) in new stack
[2026-06-10 00:20:45] VERBOSE[1476439][C-00000001] pbx.c: Spawn extension (dpma_pjsip_message_context, digium_phone_module, 7) exited non-zero on ‘Message/ast_msg_queue’

At first I thought, the phone checked config and is rebooting. But no one is pushing config changes in the middle of the night.

[2026-06-10 00:20:46] VERBOSE[1476476] res_pjsip_registrar.c: Removed contact ‘sip:[email protected]:5060;ob’ from AOR ‘26395’ due to request

[2026-06-10 00:23:15] VERBOSE[1476476] res_pjsip_registrar.c: Added contact ‘sip:[email protected]:5060;ob’ to AOR ‘26395’ with expiration of 300 seconds

2

chan_pjsip does not handle provisioning.

This tells me that the phone actively cancelled the registration, so you need to look at the phone. I suppose it might do that if it received a configuration update.

3

Specifically DPMA provisioning uses sip as some sort of convenience or security measure.

Turning on syslog seems to solve the problem… Well that sounds crazy, I believe that the phone rebooting after changing the setting is what’s actually fixing things. I’ll spend some time following up on that hunch.

4

I managed to capture some syslog messages while its happening. Right after enabling syslog, the phones were doing this “([ 538]I/dpcore: Ignoring push notification digium.config.update)”. Which in a way mitigated the problem but probably not a good long term strategy to just have your phone ignore potential config changes.

So anyway, it looks like the phone is sending registers out of order. Since both pieces involved (EPM & P-Firmware) are closed source, its hard to say what the intentions are… but the result is the phone unregisters during config check and its down until the phone sends its next periodic register request.

I can see both a register and an unregister taking place within a second:

(Bottom to Top)

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Deleting account 1..

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Acc 1: setting unregistration..

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: sip_resolve.c DNS AAAA record resolution failed: No answer record in the DNS response (PJLIB_UTIL_EDNSNOANSWERREC)

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Acc 1: Unregistration sent

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Account id 1 deleted

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/callbridge: reg_account media_encryption=

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/callbridge: Registering slot 0.0 ‘sip:pbxserver.org:5060;transport=udp’ Proxy ‘’

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Adding account: id=sip:[email protected]:5060;transport=udp

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Account sip:[email protected]:5060;transport=udp added with id 1

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Acc 1: setting registration..

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Contact for acc 1 updated: sip:[email protected]:5060;ob

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: sip_resolve.c DNS AAAA record resolution failed: No answer record in the DNS response (PJLIB_UTIL_EDNSNOANSWERREC)

Info (6)
6/16/2026, 8:48:42 AM.000
8:48:42 AM.000
[ 541]I/pjproject: pjsua_acc.c Acc 1: Registration sent

Just wondering if I’m missing something? Might be more of an official support issue than a forums issue.