vicidial.org

We are experiencing an issue with Vicidial where a call will come through to the reps phone, but there is a long delay before the fields on the screen will populate with the data from the database, roughly 15 seconds or more.

This happens randomly a few times per hour for each rep. All other calls the data and the call are timed perfectly.

Any ideas on what this could be?

We are running VERSION: 2.0.5-134 BUILD: 80703-0124. We have checked all servers to make sure they are all on the same time.

Thanks!

What's loadavg when this happens? Is this multi-server setup? Loadavg on the mysql machine would be in interest in that case.
How large vicidial_list table is in the database?
Have you checked with mtop if there is any mysql query stuck at that time?

Table size is under half a million rows, mysql has no delayed queries and mtop shows good to. This is happening at random intervals, load average on mysql is around .3-.5 and vici is at 1.3-1.5, both have 90%+ CPU idle time and more then enough free RAM.

Any ideal what else can cause this? Im looking into perl scripts now to see if the delay is there as it appears the DB is looking good.

Thanks for any help or suggestions!

Have a look at these topics:
http://www.vicidial.org/VICIDIALforum/v ... ight=delay
http://www.vicidial.org/VICIDIALforum/v ... ight=delay

I had this problem and the only thing I came up with after a solid week of researching and studying was to just reboot the dialer every night. This wipes out alot of things through the startup scripts.

The source for me was one our leads table was 13 million records and the queries that brought the data to the screen for the agents and the hopper scripts would sometimes hit right at the same time and it would just hang... after 15 to 20 seconds the data would show up and issue go away for 2-4 hours. So i just pay a tech to come in every night at midnight and reboot it. Which i hate.. Linux servers should rarely need reboots... But in dialer world (linux or windows) it seems regular reboots are in fact accepted, even daily. Sorry this doesnt help much!

How hard are you pushing it? I know the loads are low but agents, dial ratios, lines?

I did notice that if i lowered the speed of things it fell back into place as well. We would start out the day with 125 agents, all outbound campaigns using 5 to 1 dialing with answering machine detection turned on, and if it started happening (even though loads where low on all servers), i would lower dialing to 3 to 1, it would go away and I could have the dialer manager turn it back up.

v/r

KG

did you add this in your my.cnf

skip-name-resolve


had this problem before and added that to the my.cnf

or if you have that skip-name-resolve on your my.cnf

check on mtop which table is locking up.

Something definitely locking up on your mysql.

Or if not all agents experience that check the agent's pc utilization.