[SOLVED] Sometimes SIP clients are unavailable when they aren't


Over the last couple of days, I migrated from our outdated asterisk 1.4 PBX to a completely new FreePBX 15.x installation.

Unfortunately, I am still facing one major obstacle, because sometimes it is not possible to connect to (PJ)SIP based extensions, ie call them directly or use them in Ring Groups etc.

When that fails, callers are immediately directed to the extension’s VM application, even though the extension is perfectly available.

Even as I write this, calling my own extension registered on a Grandstream GXP2160 phone works at a 50% chance at best.

I’ve tried debugging the issue by maxing the verbosity and debug level on the server to 9, but as strange as it may seem, if those calls fail, there is nothing in the logs. No hint of a dialplan invoked, no hint of the VM app taking up the call etc, nothing.

If on the other hand the call succeeds, I see the all the expected things in the logs (pjsip debug stuff, dialplan debug stuff and so on).

Before I provide more details on my configuration, I am just wondering if someone has ever been in such a situation as well where the logs would not show anything and maybe has a clue how to at least turn on debugging …

With default settings, you should be able to view details about a call at Reports->Asterisk Logfiles. This will also show endpoints becoming unreachable, etc.

If only inbound calls via a trunk are failing, possibly there is a lost registration or problem with the inbound networking connection, and the call is actually being handled by provider’s voicemail, a backup PBX, etc.

Please post basic info: Cloud or on-site? If cloud, whose? If on-site, physical (what hardware) or virtual (which platform)? Do calls from another extension fail?

The logs unfortunately don’t show anything in case an extension is unavailable. I’ve also checked all other logs in /var/log/asterisk, but nothing.

In the asterisk console, I’ve set verbose and debug to 9, to no avail.

This is an on-premise installation and there is no SIP provider in between. When I hit the VM I can even make a recording and have the message sent to the configured mail address and when I cross check with our mailserver, it shows that the mail has been sent by the FreePBX server.

Outbound works flawlessly, calling from other extensions works “more often” than using inbound, so I’ve focused my testing on inbound.

Inbound is an IAX2 trunk, outbound is a SIP trunk. All devices including the FreePBX server are on the same subnet, there is nothing in between them.

well, I feel very embarrassed now, I figured out what has been going on …

Exactly as @Stewart1has pointed out, another server interfered with my setup. And it was nobody else’s server but my very own one, but the old one …

So in reality, we had been registered twice with our IAX2 trunk and thus we ended in a round-robin scenario, where calls would get answered once by our old and once by our new server. Of course, all our endpoints had already been migrated to the new server and thus our old server redirected the calls right to the VM.

Sorry for the noise :slight_smile:

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.