The 5 most popular reasons for connection failure

From years of working this board, I thought it a good idea to just post this for better understanding and trouble shooting on your own.
Hopefully this helps, but I’m open to revisions if you suggest one.

1- Lacks connection information.
Your connection data is put on a list of registered nodes.
Distributed to registered nodes.
Are you registered ? Are you in the nodes list located /var/lib/asterisk/rpt_extnodes
Does that connection info look correct in the file ? Mainly IP and IAX Port
If you are not registering and you are on a GSM Hotspot (AT&T/T-Mo) AND using asl3,
You might want to revert to IAX based registration (see asl3 manual)

2 - Do you have a path to the connection ?
The ‘dialplan’ (as asterisk names it) is the file extensions.conf
It contains how a incoming call is handled (or not)
Private nodes need routing for they are not part of the allstar system of registered nodes.
The same apply to sip and other iax connections including iaxrpt.
You qualify the connection in iax.conf or sip.conf but when you dial anything, it is extensions.conf that routes and/or connects it to something else.

3- Lack of authorization
Passwords and the like. Easily verifiable if you take the time.
Mainly when you fail to register. But includes iaxrpt or any iax/sip connection.
Allmon and other management interfaces are HTTP connections also have passwords to deal with in both apache (web access) and the asterisk manager interface (AMI).

4- A connection conflict.
Your intended target also has a private node connection as the same number.
The software prevents ‘looping’ of connections. It will be rejected.
https://wiki.allstarlink.org/wiki/Private_nodes#Numbering
But this includes if you and your target are both also connected to a particular public node.
Rejected because you are already connected, just not directly.
It can be helpful to look at a bubble map to visualize this.
You can look at the targets connections with the results of this @ links

https://stats.allstarlink.org/api/stats/<node#>

So you can look for a conflict yourself.

5- IAX Port
On inbound connection failure and your server is behind a router, the port must be forwarded to the correct IP inside your network. A improper port forwarding does not change your outbound connection ability.
Also make sure the port reported in iax.conf bindport= is corect
And is also the same port reported in the [nodes] section of rpt.conf for each node on that server if stated. (default port not required)

===

OK… there should have been a #6
If you have 2 servers behind the same NAT / router / WAN IP
There are some special considerations to be made for those connections between nodes on those servers…
This document will help.

Basically, you must describe each of those nodes in [nodes] (rpt.conf)
with the local address and port as seen from that server.

Just like they were a private node connection. Because the data from registration in the files tells it to use the wan ip to make the connection, and that will never work.
Just don’t forget to change this if you move one of them ‘outside your NAT’.

===

FAT FINGER ERRORS are sometimes hard to spot. Let some other eyes look at it.

Always test a failed connection against other nodes and see if it is widespread / all nodes, private nodes or just one node or on a particular server. That will help you localize the issue.

Do not automatically assume the problem with the connection is on your end and tinker with your settings unless you have valid idea of a incorrect setting.

3 Likes

Mike,
Thanks for the list of things to check out. We will go through the list and let you know what we find.

Dave

5 posts were split to a new topic: Other less common networking issues for ASL