Multiple agetns fall in the same Session ID simultaneously

[p_amador]

Hi Everybody,

I am trying Vicidialnow CE 1.3.

Sometimes two or three agents fall in the same vicidial conference when they login and obviously can hear the other one talking. I workarround deleting the vicidial conference and login again the agents in conflict one by one, and later adding the conference deleted for future use. Please, if someone experiencing the same problem and solve it, share the solution.

Is possible to tie a vicidial conference with a phone and each phone can use always the same session ID to prevent this issue?

Thanks,
Best Regards.

[p_amador]

Guys, I read all the posts of the topic propossed by "globalme" for a similar problem, and it looks like a nightmare to solve, and finallly he goes to a fresh instalation, but no more posts after it. In my opinion, is clear, that only happens when two agents share the same conference room (session ID). Search the reason for that is the really way to solve it, check the carrier or the internet conection only spent the time, the processor load maybe, but he says that occurs with few agents. This reading help me to see another troubleshoot technique to review tomorrow with my agents, the propper way to managing the interface. I really like to know if a similar problem was posted and solved with a particular recommendation.

Thanks.

[gardo]

Here are some of the reasons:

1. System load too high
2. Misconfiguration
3. Agent not logging out properly or stale sessions
4. Mysql data corruption

What's your hardware spec? How many agents dialing? Asterisk version? Vicidial version?

[p_amador]

Thanks gardo, the answers:

Here are some of the reasons:

1. System load too high (maybe but I dont think so, it occurs early in the morning too, with few agents connected)
2. Misconfiguration (I did not touch the code manually, vicidialnow comes with many extensions and conference rooms created by default, it is very useful)
3. Agent not logging out properly or stale sessions (maybe you have a point, I work in training my agents to prevent mistakes, but obviously is possible, and I wish it to be the cause of the problem, because is easy to solve)
4. Mysql data corruption (Auch!!, it could be a really problem, you mean the database not register the session was ocupied by an extension and reuse the session?

What's your hardware spec?
DELL Poweredge 2950, Xeon QuadCore 2.8 GHz, 2 GB RAM, 140 GB HDD
How many agents dialing?
My primary campaign is inbound, it could be 25 agents logged in in a heavy work hour.
Asterisk version?
1.2.30
Vicidial version?
VicidialNow CE 1.3
In system settings Version: 2.0.5 DB Schema Version: 1135

If could be possible to tie a Vicidial Conference Room with an extension, it may prevents this problem. If you have an Idea to go in this direction, please tell me, I could investigate and share with the community my job.

Thanks again.

[gardo]

Your hardware is powerful enough to easily handle 25 inbound agents. So we can safely disregard having too high a system load. Though it will still help if you can post the output of "top -c" when you're having multiple lines in a session.

Try this quick fix:

Manually edit /etc/asterisk/sip-vicidial.conf. Replace all the "context" entries on your phone extensions from:

[cc100]
username=cc100
secret=test
mailbox=100
context=default
type=friend
host=dynamic

to

[cc100]
username=cc100
secret=test
mailbox=100
context=nocalls
type=friend
host=dynamic

[williamconley]

never safely assume system load is not high. always check system load when "irregularities" occur. after all, other things can be happening that can hammer the system, and you may be able to backtrack through top or htop to find WHAT.

unless you don't get the system load.

please post your entire configuration

vicidialnow is your installation method, i presume, but we need to have your vicidial version and build, whether there is any telephony hardware installed in your box, asterisk version, cluster setup if applicable, and whether there is any OTHER software installed in the box.

i would recommend posting your server load from top or uptime or any one of several other locations (or observe htop during the time this happens and See if you are maxing out).