FreePBX Endless Loop

[tshif]

This doesn't belong in the PBXIAF forums, it belongs in the FreePBX forums. And it is. It was Posted by PBXIAF forum friend TheShniz LockUp when calling Unregistered Extension

It was posted WAY back in Jan 0f 2008. No one in the FreePBX group has ever responded in any way. THAT'S why I'm posting it here. Usually, someone RESPONDS here.

I can not speak to all the point raised in that thread, but I will say that without a doubt, this statement is true. I can duplicate it on every PBX we have installed, and the two we inherited from other service providers.

1.) (calling a valid, but unregistered extension, and then) hanging up & not leaving a message during the 'please leave a message prompt' causes loop/lockup

I will add to that the fact that it doesn't matter weather you call from extension to extension, or via a DID, the same results occur.

This happens on the most current version of FreePBX, 2.5.1.1. It seemed to have begun around the time VMX was introduced to FreePBX. (Can VMX be uninstalled by itself? We NEVER use it, or configure it and I am suspicious in that the endless loop ALWAYS references it)

Sorry if I sound a little terse.
I guess I'm a little frustrated by the long term nature of this issue, the easy reproducibility of the problem, and mostly, the way the folks at FreePBX have IGNORED the posts in their own forums about it since January of 2008. Maybe they really need to have a forum category for BUGS so their team can find them easier. I wont be wasting my time and posting a bug in their bug tracker. The ever present response of "Closed: Works for Me" is just to irritating - and reminds me that I wasted my time to even create the bug entry.

Since this issue has been around since FreePBX 2.4 - someone MUST have found a resolution to it. If you have, would you be kind enough to share it with us? TheShniz, did you ever find a way to prevent this?

tshif

 

[p_lindheimer]

Is there a bug filed on FreePBX.org and if so can you reference the bug number. I recall seeing this reported and could never reproduce it, and I just tired to reproduce it un-successfully now.

There is not bug topic in the forums because there is a bug tracker. Post a bug or bounce the existing one if there is one, and provide very detailed repro instructions including the extension configuration, logs, etc. - since it appears from the test I just ran that without that we may not be able to repro it.

 

[GlenPittman]

We were seeing this happen on our boxes that were PIAF1.2 (Asterisk was version 1.4.21.2) but not on our PIAF 1.3 installs using Asterisk 1.4.21.2.

I could never recreate this, but I do have a hard drive with a load that was producing this about twice a week. I can put it back into another box for testing, but I will have to find a Digium PRI card for testing. We use Sangoma now (for better results using faxing). I do have a few PIAF1.3 with Digium cards and they do not loop, so the change to Sangoma wasn't the cure.

Our PIAF1.2 boxes used IAXModem with Hylafax and Avantfax. We had the pager notification script installed, and also had the Aastra scripts from Ward loaded. We are not using the IAXModem, Hylafax, or Avantfax in PIAF1.3

We had the issues on boxes running all Polycom phones, and all Aastra Phones. We have had the loop on systems with Analog Trunking and with PRI trunking. None of our boxes were using SIP trunking or IAX trunking.

I can't think of anything else right now that might help shed light on what might have been the issue.

I do have one box out there running PIAF1.2 with Asterisk 1.4.18.1 and I do not have the issue with it. It did have IAXModem, Hylafax, and Avantfax on it.

Glen Pittman

 

[tshif]

Ok - I'll do it. It'll take a day or so before I can make the time for a detailed procedure to reproduce - Which logs specifically do you desire?

 

[p_lindheimer]

Enough detail from logs or CLI traces to see exactly what is looping, and enough of the log leading up to the issue to see how it got there.

 

[tomsyr]

Getting support

This reminds me of a time when I had a problem I just couldn’t figure out. If I didn’t have a PIN code in an outbound route, calls would not go through. If I put a PIN code in (any number) the calls would go though! I discovered this ‘solution’ by trying to see where the failure was. It made no sense, and no other system I worked on behaved that way. I finally broke down and paid a developer to ssh into the box to help me solve it. A bug was found and corrected within 15 minutes. The point I’m trying to make is that sometimes a little paid help is worth it. Now, if I could only get back the 10 or so hours I put into trying to figure it out myself

 

[tshif]

I need to make "double-dumb-ass-on-me" sure I can reproduce it as easily as I think I can. The last thing I want to do is waist someones time in a dumber-than-average way.

p_lindheimer has a well earned reputation of listening to, and working with, anyone who can bring a reasonable concern and valid data for reproducibility to the table.

I have just prepared the test-beast box - and will begin documenting all steps taken to arrive at the problem, and then post or bump this on the FreePBX bug tracker.

I really hope its me and my configuration, or a bug they can find, cause its really a bummer to deal with it.

 

[TheShniz]

Sorry guys, I've been a bit backed up w/ simultaneous installs...

To give some background/quick-facts on my original posts:

1.) This took place a month or so after PiaF was first introduced
2.) I was using TB 2.4 at the time
3.) TB 2.4 marked my personal transition from Asterisk 1.2 to 1.4
4.) I was working as hard/fast as I could to get new installs off TB, and onto PiaF... but alas, systems had to get out, and I had to go with what I already had working.

Here's some age-old lessons:

1.) When a modification to something has the ability to impact its stability/performance (however unlikely), it is always best to apply said modification alone, without additional modifications, and to confirm results with testing. The same is true for automobiles as it is telephone systems.
2.) Digium and backward compatibility should never be used in the same sentence, unless it is describing the complete lack thereof.
3.) Assumption is the mother of all... fubar's.

In essence, too much was happening all at the same time, as I was transitioning from Asterisk 1.2 to Asterisk 1.4, and again from TB to PiaF (as fast as humanly possible).

After all is said and done, and more was said than done... my problems were created by transcribing & implementing BobH's VM Notify Script into my 1.4 base, with a very small yet significant syntax error (I missed one), as result of transitioning to Asterisk 1.4. It should be noted, that I've used his script since its conception throughout Asterisk 1.2, and knew it to be rock solid.

It was most obvious, that the problem resides squarely in the voicemail application (somewhere), and I too thought my problems were that of VmX Locator (still pretty new at the time).

Assuming the above is also your problem, I see one of the following possibilities:
1.) [macro-vm] has changed, but you've not updated VM Notify
2.) VM Notify has syntax errors
3.) Digium has changed syntax again (unlikely)
4.) VmX Locator is fubar (unlikely)

I would suggest you remove all custom scripts, and add them back one at a time... I suspect you'll find the culprit there.

Sorry for the novel,

::updated::

Original Posts:
http://www.trixbox.org/forums/trixb...ed-lockup-when-calling-unregistered-extension
http://freepbx.org/forum/freepbx/users/lockup-when-calling-unregistered-extension

I purposely use an email addy to register in the various forums that I do not check, in an effort to cut down own the amount of spam I receive... as such, I don't (usually) get forum reply notifications. I've made a quick reply to the FreePBX forums, which I should have done originally, sorry.

 

[tshif]

Excellent!

And BINGO besides! That script WAS present, but in a conf file I rarely if ever change (or obviously look at). I am delighted to find I caused my own problem - makes the fix so much easier. Haha.

TheShniz - thanks for the solution - I might never have thought to look for that script where I had it running from.

p_lindheimer - I was about to have to admit that I could not so easily reproduce the problem on a new naked install. So I'll have all my crow in one sitting this time - and thank you for always being ready to look and listen.

I hope I can discover a way to correct the problems with that script - its very useful. Does anyone have a version of it running that doesn't cause the dreaded loop of death?
:hangb:

 

[TheShniz]

It *really* is a great script, and an excellent workhorse to be sure... I've secretly hoped for quite a long while that Phillipe and the fellas might integrate it into FreePBX for mass consumption, plus it would also simplify things w/ [macro-vm].

As requested, here's my prev post w/ the code:
http://pbxinaflash.com/community/threads/emergency-after-hours-pager-script.1909/?t=1909

I'm glad you got things patched up Tony, and hopefully Glen won't be too far behind!!



To reiterate:

Be sure to update [macro-vm] from /etc/asterisk/extensions.conf as it has been known to change from time-to-time (which can *really* fubar things up if you're not paying attention).

That's it folks! I should add this script is an excellent display of open collaboration, is tried & tested, and you may find a good many customizations/localizations/etc already available (see prev post).

Special Note:
My previous post may or may not contain the latest [macro-vm], it would be wise to double-check