Customer hangup not detected in asterisk

[jkidd]

Hi, we've just installed a new cluster, everything is working well but there is just one issue that the customer hang up is not being detected, so the agent doesn't realise if the call has hung up.

On our previous system, asterisk log will show the following if the customer hangs up the phone;

[Feb 18 08:37:19] == Spawn extension (default, 8600051, 1) exited non-zero on 'SIP/PC010172-00003ff5'
[Feb 18 08:37:19] -- Executing [h@default:1] AGI("SIP/PC010172-00003ff5", "agi://127.0.0.1:4577/call_log--HVcauses--PRI-----NODEBUG-----0---------------") in new stack
[Feb 18 08:37:19] -- <SIP/PC010172-00003ff5>AGI Script agi://127.0.0.1:4577/call_log--HVcauses ... ---------- completed, returning 0

On our new system, we see no logs appear when the customer hangs up. We're sure this isn't the carrier as we're using the same SIP Trunk.

One thing I notice is this warning message which didn't used to appear on the old system, some function related to determining the hangup cause;

[Feb 18 09:34:37] WARNING[3428][C-0000000c]: func_hangupcause.c:140 hangupcause_read: Unable to find information for channel

Thanks for any help!

VERSION: 2.14-697a
BUILD: 190121-2019
Revision: 3063

1 DB server, 4 Web / Diallers

[williamconley]

1) Please list your installation method With Version during all posts. It makes a difference.

2) Did you import a database from a prior system?

3) Post your Asterisk Versions from:

Code: Select all

grep asterisk_version /etc/astguiclient.conf -in;
select server_id,asterisk_version from servers;
asterisk -V;


[jkidd]

Thanks for the quick reply William

1-Installation method was from the vicibox manual and the pound team guide - basically built 5 servers, updated them to the latest svn and run vicibox-install on DB followed by 4 servers that all share dialler / web services
http://www.poundteam.com/downloads/Vici ... v1%201.pdf

2-No data imports

3-
asterisk_version=>13

server_id asterisk_version
harryDB 13.21.1-vici
harryWD02 13.21.1-vici
harryWD03 13.21.1-vici
harryWD04 13.21.1-vici
harryWD01 13.21.1-vici

Asterisk 13.21.1-vici

[williamconley]

installation method should include the Installer with Version (eg: Vicibox 8.1.2 ...?)

Have you altered any of the .conf files manually?

Please also post the dialplan entry from your Carrier.

[jkidd]

Installer was ViciBox_v8.x86_64-8.0.1.iso

Only alterations are on astguiclient.conf

Changes:
 VARactive_keepalives => 123468 (on all but one server)
 VARactive_keepalives => 12345678 (on one server only)

Dial Plan;
exten => _4XXXXXXXXXXX,1,AGI(agi://127.0.0.1:4577/call_log)
exten => _4XXXXXXXXXXX,2,Dial(${SIPTRUNK4}/${EXTEN:1},,tTo)
exten => _4XXXXXXXXXXX,3,Hangup

[williamconley]

not seeing anything out of the ordinary.

i guess now we need to see evidence of your flaw.

and of course, be sure you mean "asterisk" and not "vicidial" as the package not detecting.

no activity when a "customer" hangs up could also be interprested different ways. what kind of call is this? manual? autodial? inbound?

have you compared to some other server's cli to find the technical difference between this "broken" and that "working" on a test call for each (no other traffic, just that test call, so there's not any extraneous information).

sip debugging can also be used to deep check. if there's no different there, agi debugging may offer some information.

[jkidd]

Thanks William,

The test I did was to watch the asterisk log (asterisk -vrrrrrrrrrrrrrrrrrrrrrrrr)

On the working system, if I (as the customer hang up) I see the following;

Code: Select all

[Feb 18 08:37:19] == Spawn extension (default, 8600051, 1) exited non-zero on 'SIP/PC010172-00003ff5'
[Feb 18 08:37:19] -- Executing [h@default:1] AGI("SIP/PC010172-00003ff5", "agi://127.0.0.1:4577/call_log--HVcauses--PRI-----NODEBUG-----0---------------") in new stack
[Feb 18 08:37:19] -- <SIP/PC010172-00003ff5>AGI Script agi://127.0.0.1:4577/call_log--HVcauses ... ---------- completed, returning 0

On the new system, if I (as the customer hang up) I see nothing in the log. This shows as a symptom to the agents as they don't here a beep to indicate the call has terminated.

How do I do sip debugging? Thanks for the help

[williamconley]

what kind of call is this? manual? autodial? inbound?

sip debug is in this along with many other places https://www.voip-info.org/asterisk-cli/

[jkidd]

Sorry, the call type is autodial, just on ratio mode

[williamconley]

does the system recognize hangup on other call types?

is your other server using asterisk 13 also? (the one you claim validates your carrier, of course)

does your other server (if using asterisk 13) use the same sip.conf account settings for this carrier?

[jkidd]

Just tested on manual and it doesn't detect it. Agent web UI still says LIVE CALL (it does detect the call is dead after about a minute)

The other system is much older, SVN 2590, asterisk version 11.23.1-vici (single server)

I compared the sip.conf files, the only difference between old and new is the external IP we're displaying (externip = 10.xxxxxxx); Address that we're going to put in outbound SIP

Using SIP debug, I'm not even really sure what to look for. Do you think it has anything to do with this warning in asterisk log?

Code: Select all

[Feb 18 09:34:37] WARNING[3428][C-0000000c]: func_hangupcause.c:140 hangupcause_read: Unable to find information for channel

[williamconley]

Just tested on manual and it doesn't detect it. Agent web UI still says LIVE CALL (it does detect the call is dead after about a minute)

i'm not talking about the vicidial agent interface, there's a lot of lines of code between hangup and that screen noticing.

i'm talking about a manual dialed call. does asterisk terminate the call or leave it open?

[jkidd]

Thanks William, asterisk leaves the call open

[williamconley]

sounds suspiciously like asterisk is not using the correct sip.conf base file to me. but you'll need to use sip debug to confirm that a hangup command is in fact being received. if it is NOT, then your assertion that the carrier is not at fault is "questionable".

did you have any "do-over" moments during the installation? perhaps have to change the asterisk version number?

Also: If you just finished installing this with 8.0.1, how did you manage asterisk 13? If I recall correctly, that installer is for asterisk 11 and 8.1.2 is for asterisk 13.

Consider a perl install.pl (which doesn't ever hurt) again and verify the asterisk version during the install process. Note that all this process does is copy all the scripts where they belong.

While performing this, be sure to use --copy_sample_conf_files which will copy all the base .conf files again. Any that have been edited will need to be re-edited (but there should only be ONE edit: sip.conf's externip value). It will remember your ftp or other values in /etc/astguiclient.conf, but most importantly if provided with the correct asterisk version (from the provided list, don't make up something that is not in the list!! if it says "11.X" use "11.X" not "11", for instance).

[jkidd]

Hi William, thanks again for the help.

Turns our you are right, the error was in sip.conf on the diallers and it's an issue with the telco (/etc/asterisk/)

Our external IP (the only thing that was different from the last server) had the wrong external externip address for our sip provider. I think this means when they're trying to send information back to us like the customer hanging up, it was going to the wrong place and not being detected.. so yes Telco related (sorry..)

Thought I'd post an update in case others are reading this info - essentially if you are seeing no response from the customer pressing hang up, even in sip debug, it might be telco related.

Thanks!

[williamconley]

good postback.

and you learned that nothing proves something isn't the error until all variables have been accounted for. what's left ... is the problem.