Quoting from the documentation AND the module page:
[QUOTE]Note that Asterisk is incapable of having two identical routes and trying to force calls to use the other route if one of them is banned by this module. It will not work. You must have unique outbound routes for the proper selection to work
If you wish to emulate this functionality, you can use the ‘Redirect’ function. …[/QUOTE]
Note that just having a different trunk selection does NOT make the routes non-identical!
Perhaps it will help if I explain it this way. When you place a call, irregardless of what you may or may not have done in routepermissions, Asterisk goes through your routes one by one and tries to match the number called to the patterns in your routes. It stops searching on the FIRST match it finds, and that’s it - under no circumstances will it look at any other route once it’s found a match. Only AFTER it has found a match (actually, only after it’s already sent the call to a trunk) does it check to see if the user has permission to use that route. If yes, the call goes through, but if no, the call stops dead in its tracks.
Let’s say you have a second route with identical dial patterns as the first. Your outbound calls will never use it, no matter what you do. Remember: Asterisk stops searching on the FIRST match, and it doesn’t check to see if the user is allowed to use the route until AFTER it’s made that match.
So that’s the point of the redirect prefix. Let’s say your first route has the pattern 1NXXNXXXXXX (not something I’d recommend unless you want to allow some really high-cost calls to the Caribbean, but it’s just an example here). If in your second route you also put 1NXXNXXXXXX, that pattern will never be matched. It HAS to be unique. So what I might do is instead use something like 0001|1NXXNXXXXXX for the second route. Then when you deny access to that first route, you put the 0001 prefix in the “redirect” text entry box. Now let’s say you make a call to 1-800-555-1212 from an “alternate route” extension:
● User dials 1-800-555-1212
● 18005551212 is sent to from-internal context which begins looking for a match on the number in the route dial patterns.
● A match for 18005551212 is found in a route, the one and only route that will ever be used for the number 18005551212.
● The call is then sent to the first trunk in the list associated with that route.
● One of the first things the trunk does is to determine if the user (identified by Caller ID number) is allowed to place calls via the route that the call just came from (which is still available in a variable). In this case it finds that no, the user is NOT allowed to place a call on this route, BUT that a redirect prefix of 0001 has been supplied
● The called number is then modified to be 000118005551212
● 000118005551212 is sent back to the from-internal context which begins looking for a match on that number in the route dial patterns.
● A match for 000118005551212 is found in a route (hopefully NOT the same one that would match 18005551212), the one and only route that will ever be used for the number 000118005551212.
● Because of the bar character in the dial pattern, the digits 0001 are removed from the called number - note that at this point the route has already been selected - so the number again becomes 18005551212 before being passed to the trunk.
● The call is then sent to the first trunk in the list associated with that route.
● One of the first things the trunk does is to determine if the user (identified by Caller ID number) is allowed to place calls via the route that the call just came from (which is still available in a variable). In this case it finds that yes, the user IS allowed to place a call on this route.
● The call then goes out the selected trunk - or if that trunk is busy, it will try any other trunks associated with that route.
If you are truly following along up to this point, you might realize there’s a slight inefficiency here if a trunk is busy, because the user’s permission to place the call is checked each time the use of another trunk is attempted. However as a practical matter you’d never notice the additional time because the check is very quick.
I hope that helps you understand why you can’t use the same pattern in two different routes and expect it to work. You must use a redirect prefix on at least one of the patterns so that it will be recognized as unique.
Off-topic comment: I know how some of you must feel when I’m rattling off these explanations and maybe you’re just not getting it. I’ve been in the same position the last couple of days, trying to set up a VPN using OpenVPN, and most of the explanations I see assume a whole lot of knowledge that I don’t have, and therefore the instructions make absolutely no sense whatsoever. I try to simplify some of these explanations as best I can, and I’m truly sorry if some of them are still “clear as mud.”
