So, we migrated a 2.11 system with about 100 phones ( which worked perfectly for many years) to FreePBX 17 via a backup and restore. Following the migration, things mostly worked- some did not. But we got through the little issues. Calls, both inbound and out, paging, email notifications, all working. But not completely tested…
Once all looked good, (but still not completely tested,) we swapped IP addresses to make the new version 17 the live, production PBX. After a few hours, reports started coming in that inbound calls would drop at 30 seconds. Ah, that should be easy- some NAT setting somewhere. But the cause of the drops weren’t obvious. After half a day of struggling, we reverted to the 2.11 version.
I probably should have studied this more before making the change. But I still don’t understand where I went wrong.
The PBX is natted in a data center. The network settings appear to have migrated correctly- and the autodetect seems to correctly populate the Exterenal address. We have an SBC which the endpoints all point to. Inbound calls point to a seperate PBX on the same LAN subnet, and we use IAX2 trunks for inbound and outbound calls. I’ve tried both IAX2 and SIP trunks- both fail at 30 seconds. There is two-way audio until the drop.
FreePBX 17 (and the current version of Asterisk) is pretty unfamiliar to me. So possibly I’m missing something very basic- like a check box somewhere. The firewall rules have not changed. There is no SIP ALG or SIP Helper on our PfSense firewall.
So, it was a working FreePBX 2.11 → migrated to FreePBX 17, same IP address, no changes to the trunks, etc… and now calls drop at 30 seconds.
I suppose an Ethereal or Wireshark trace is next- but maybe I missed something simple, somewhere?
Did you restart Asterisk after updating the IP address? Even though it shows the correct IP in Asterisk settings you still need to restart Asterisk for it to work properly.
I think you’re ignoring the bigger setup picture. The PBX in question was migrated to FreePBX v17 but there’s an SBC between that system and the users. The PBX, itself, is behind NAT so now there’s two NAT’d sides which is bad. There’s also a separate PBX that inbound calls get routed too and finally they are using IAX in between things.
This doesn’t seem to be a PBX that is directly connected to upstreams or the end user’s phones but connects to a bunch of other things they should make sure work properly. So this can’t be treated as ITSP directly to the PBX or phones directly on the PBX. There’s stuff in those paths that handle the calls.
When this was all setup years ago, (when FreePBX 2.11 was new!) we had our current topology. And there is still another 2.11 system in this same natted LAN topology. Everything worked well- and the remaining 2.11 system works fine.
But clearly something has changed. Not with the topology, or our physical firewall rules.
Yes, you upgraded to a PBX with newer OS, software and all sorts of changes from the last 15 years. You had chan_sip and that’s gone now so now you have a completely different SIP driver to contend with. IAX has had changes…plus you have all these other systems that connect to it…you need to make sure those connections are still working as expected.