vicidial.org

Hi,

We are running a vicidial cluster installed with ViciBox, with a database, 4 dialers a replication database and a web server.
VERSION: 2.12-485a
BUILD: 150512-2225
We can successfully make calls with manual dial or by manually dialing a call. As soon as the destination starts ringing the live call indication on the agent screen changes to call hang-up. This is a false indication because the call is still alive and working fine.

On the real-time report screen the table for realtime statuses opens when an agent log in, but doesn't have any rows showing active agents. No live information is shown when an agent makes a call:

DIAL LEVEL: 0.000 TRUNK SHORT/FILL: / 0 FILTER: NONE TIME: 2015-05-18 09:10:47
DIALABLE LEADS: 10086 CALLS TODAY: 51 AVG AGENTS: 0.00 DIAL METHOD: MANUAL
HOPPER ( min/auto ): 200 / 0 DROPPED / ANSWERED: 0 / 4 DL DIFF: 0.00 STATUSES: NP, N, ERI, DROP, DC, B, A, NEW
LEADS IN HOPPER: 391 DROPPED PERCENT: 0% DIFF: 0.00% ORDER: RANDOM
NO LIVE CALLS WAITING
0 agents logged in 0 agents in calls 0 agents waiting 0 paused agents 0 agents in dead calls 0 agents in dispo

VICIDIAL: Agents Time On Calls Campaign: |04| 2015-05-18 09:10:47
+----------------+------------------------+-----------+----------+---------+------------+-------+------+------------------
| STATION | USER SHOW ID INFO | SESSIONID | STATUS | MM:SS | CAMPAIGN | CALLS | HOLD | IN-GROUP
+----------------+------------------------+-----------+----------+---------+------------+-------+------+------------------
+----------------+------------------------+-----------+----------+---------+------------+-------+------+------------------


We have verified all the VARactive_keepalives settings in all the dialers. We have also verified the cron-jobs and we have search though the log files for errors, but we could not find any.

The admin user has got all the rights allocated.

Any ideas would be really appreciated.

A time sync issue comes to mind, make sure all those servers have the same time down to the second. If they don't setup a quick NTP server for all your servers and clients to use.

Other potential issues might be database interactions, altho I would expect more violent results. Still worth looking into. Are all those servers sharing a subnet or could there be interference with some of them (by, say , a firewall appliance or a fancy router?). See if all of them can reach MySQL with good latency, bandwidth and signal quality.

Regarding the keepalive flags, you'll also want to check those from the DB and web server. The sample file that comes with Vicibox should included comments on what flag to enable on what server type. Once you checked the flags on any given machine, run ADMIN_keepalive_all.pl with debugX enabled to make sure it's all going fine. If you do have DB interaction issues, they'll show up here (among many other potential red flags).

Hi, I had a look at all the things mentioned, but I just cannot find the issue. This site has two Vicidial clusters. I managed to get one working (accidently I suppose because I cannot replicate it) , but the other will not work even with the same configuration. I have even done an installation with a newly downloaded iso with the same results. The installation is done by the book. The configurations are the same of another customer's Vicidial cluster. I have done tests from the problem sight connecting thought to the another customer's Vicidial cluster and it works, so the problem is not specific to the site and it's networks. I have installed a couple of clusters before that has being in operation for more than a year without issues. I just can't understand how this problem can occur from a new implementation with nothing configured more than to make a single manual dial test.

We're going to need more info. What's the output of the Asterisk console during one such call? What about the AGI logs? There's a hint somewhere, just gotta find it!

The problem occurs before a call is made. The realtime report doesn't show an agent when the agent logs in, but show an empty row.

I have also noticed that the TIME field in the Real-Time report show the incorrect time (it is sex hours behind, my time is GMT + 2). I made sure that all the GMT settings that I know off are set correctly and that all the server's times are in sync.