Got a network-related issue with Asterisk/Vicidial


#1

Hey all,

We have about 12 users that are using Vicidial for some outbound calling campaigns. The machine is a single proc 3.0 P4, with 2GB of RAM, running fedora core 4, asterisk 1.2.4, zaptel 1.2.5, and astguiclient/vicidial 1.1.9

Agents are connected via similar machines (2.8 P4s, 256 or 512MB of RAM, 10/100 e-net) and are running eyebeam as a softphone, with firefox as the dialer client, and our legacy accounting app. Nothing else is open.

We’ve been having issues off and on with dropped calls and some audio delay, and I was finally able to actually hear a call drop, as well as listen in on what the delay is like.

In researching this, I’ve done just about everything under the sun, but for some reason, I decided to do a netstat -n on teh dialer box. Needless to say, i was suprised when nearly 800 lines showed up. Here is a small sampling:

tcp 0 0 127.0.0.1:3306 127.0.0.1:46330 TIME_WAIT tcp 0 0 127.0.0.1:3306 127.0.0.1:46325 TIME_WAIT tcp 0 0 127.0.0.1:3306 127.0.0.1:46326 TIME_WAIT tcp 0 0 127.0.0.1:3306 127.0.0.1:46323 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4147 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4146 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4145 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4144 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4151 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4150 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4149 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4148 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4155 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4154 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4153 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4152 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4157 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4156 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4131 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4130 TIME_WAIT tcp 0 0 ::ffff:10.100.6.60:80 ::ffff:10.100.2.244:4129 TIME_WAIT

Most of the 800 lines are simlar to the lower half of the above snippet - there are literally a hundred open connections for each person on the dialer. This doesn’t seem right. For one of our production servers, with almost three times the number of users, there are fewer than 100 entries in netstat.

I basically want to check and see if this is to be expected (assuming Matt Florrell will have to chime in on this one) and what I might do to try and troubleshoot it. I’m not a network guy by trade, and will be researching this, but figured I’d ask here as well.

Thanks, and please let me know if you have any suggestions.

Wes