Anyone have any ideas on how to solve an issue with trunks NOT automatically returning to registering after an outage has ben restored, even after a system reboot?
I am having an intermittent issue which is ongoing and causing issues. Does anyone have configuration suggestions (or other suggestions) that could alleviate the issue?
I am on Wazo 23.10 on a Provmox VM. My 5 trunks are all with Vitelity. I have a fixed external IP address for my location and fixed internal IP address for the NATted Wazo server in my location.
This configuraiton has been in place for over a year and normally works very well for my small home/office installation with 8 landline IP Phones and a few other softphones occasionally connected.
The problem is when I lose internet connection, which can happen about once a month. The trunks connected to Vitelity lose registration when the internet connection goes down but if the internet outage is not a quick blip (ie lasts, say, 5 minutes or more), when the interent is restored, the trunks connected to Vitelity do NOT come back. Instead, when I enquire with
asterisk -rx "pjsip show registrations"
I get back, for each of the 5 Vitelity trunks a status of rejected.
The trunks will NOT re-register by themselves and will NOT re-register even if I reboot the system.
I need to shutdown the system and wait at least 15 minutes to start the system and then all is well again.
So it sounds like a firewall/fail2ban issue but I canot figure out what firewall/fail2ban issue would cause this, especially since a temporary (15 minute) shutdown and startup restores all to normal.
To say the least, not a good scenario. And since My “users” are my wife and daughter (and my clients and our friends who call), having them unhappy is not a good thing.
Unfortunately/Fortunately, the issue intermittent and troubleshooting/testing solutions is not easy to manage so I am asking if anyone has seen a similar issue and if so, what was their solution?
My trunk registration with Vitelity is set in a Template for all trunks and includes, in the Endpoint settings:
identify_by username,auth_username,ip
I include this bit if information since I read somewhere that NOT using IP registration is a potential issue but I want to be able to differentiate which trunks are in use so would prefer to keep the identify_by setting as is but if it is causing the issue, then I will change.
I also read somewhere that if the provider sends a 403 (instead of 401) response to a lost registration this could cause the issue and in chan_sip there was a setting called
register_retry_403=yes
but I cannot find an equivalent in chan_pjsip nor would I know where to enter it using the Wazo GUI (which is what I use for all configurations)
So, I have exhausted my technical knowledge.
Anyone have any ideas on how to solve the issue of trunks NOT automatically returning to registering after an outage has been restored, even after a system reboot?