Agents going to DEAD status

[ivschakravarthi]

Hi,
I have an Inbound Callcenter and Sometimes(may be once or twice in a day) suddenly calls are not routing to the agents and when i check the real time report,i see all the agents in the DEAD state.

Again after 1 or 2 mins,every thing is normal.
I checked and there are no logs of asterisk getting restarted at that time.

How can i trouble shoot it?
Please help



Vicdial VERSION: 2.8-403a |BUILD: 130510-1350 |asterisk 1.4.27.1|single server |Centos6.5 |One Digium card | No cluster

[striker]

hi
as per my understanding
The Dead Status in realtime report means that your agent have got a call and the customer have Hanged up but still your Agent didnt pressed the Hangup Customer Button to move on to Disposition screen.
also you can check the agent performance report for the time spent in DEAD status.

[ivschakravarthi]

But it is a situation,where all the agents are going to dead status at a time and retrieving back at once.

[GaD]

This could be cased by different things. What is happening is that your calls are getting disconnected at once, hence all agents are in DEAD state. Log on to asterisk's console and set a reasonable verbose level and try to catch the moment where the calls get disconnected, check why they are being disconnected, if it is upstream (to your carrier -eg. lost connection to your carrier) or downstream (your agents became unrachable to the server). Lets start there, post your findings.

[ivschakravarthi]

I am getting following mysql error on the agent screen when they are in dead state in real time report.

my SQL connect Error:can't connect to local MY SQL server through socket"vallib/MySQL/MySQL.socket"(11)

When i looked in to mysql for processlist, vicidial_agent_log table has been locked by a select query and other queries are waiting for the lock to release.
As per my observation,
2 select queries locked vicidial_agent_log table one after the other immediately

1. | 772129238 | cron | localhost | asterisk | Query | 56 | Copying to tmp table | select full_name,vicidial_users.user,sum(pause_sec),sub_status,sum(wait_sec + talk_sec + dispo_sec)


2.| 772135256 | cron | localhost | asterisk | Query | 63 | Copying to tmp table | select count(*) as calls,sum(talk_sec) as talk,full_name,vicidial_users.user,sum(pause_sec),sum(wait

I think these queries are coming from Agent performance report page.

I used to get around 10000 calls per day and only 5 months data in vicidial_agent_log table .
I cleared the table today in the early morning when they are no calls.

[ivschakravarthi]

After clearing our the tables also , i am still getting the mysql error now and then on my agent screens.

my SQL connect Error:can't connect to local MY SQL server through socket"vallib/MySQL/MySQL.socket"(11).
What else i need to do?