Calls "hanging" up svn 2.0.5 188

[telconetcol]

Sometimes when the agent is in the middle of the call (version trunk svn 2.0.5 188) appears a green display saying go back or call agent again.

The audio continues despite of having that display. Finally when the customer or the agent hangs up, there is no way for having the dispo window.

[mflorell]

What kind of calling are you doing?

I don't understand the "there is no way for having the dispo window". What shows up on the agent screen?

[telconetcol]

Manual dial with lead preview.

The call continues but in agent screen appears a green display that says:

Call agnet again
Go back

If the aqgent say call agent again, the call actually hangs up.

So the agent continues till the customer hangs up... and after press call agent again, cause for the system there is no agent in session. No way for having call dispo.

[mflorell]

Asterisk version?

loadavg when this happens?

[telconetcol]

Asterisk 1.2.17.

load: 42%
mem = 80%

[mflorell]

is that 0.42 or 42%

[telconetcol]

42%

[mflorell]

42.00 is way too high, you need to lower the load.

[telconetcol]

ok

any sugestions for achieving it??

i have got model name : Intel(R) Xeon(R) CPU 5110 @ 1.60GHz
but only with 5 agents taking calls through a sip trunk.

Thanks for your reply!!

[williamconley]

5 agents is good information. but what about the rest? how many simultaneous calls? what kind of calling?

example:
5 agents logged on via SIP soft phones using ulaw
15 g729 calls in a "press one" broadcast campaign.

this will give more of an idea.

5 agents seems small, if you are running 1:1 dialing. but if you are running 20:1 voice broadcast ... that's a different story entirely.

and can you post the actual stats for the load on the server, that may help as well.

[telconetcol]

ok... ok...

5 agents calling in manual mode with preview.

all of the calls go through sip trunk through internet. no broadcast.

have got 1G ram. Used normally 80% ram

the processor sometimes rise up till 45%.

Thanks for your reply.

[telconetcol]

ahhhh all of the agents are using ulaw codec.

the sip trunk is g729.

[williamconley]

how do you know the cpu goes to 45%? can you post the message that tells you this?

what other software are you running? (I mean besides the software automatically installed in vicidialnow or scratch install)

you are in asterisk 1.2.17, vicidial svn 2.0.5 188

Is this vicidialNOW or scratch install on your own (which OS)?

I have had issues a couple times with "transplanted" boxes where the CPU did not match the compiled asterisk and recompiling asterisk resolved this, but that's just a "shot at the dartboard" approach.

[telconetcol]

Looking at top.

top - 09:18:44 up 1 day, 9:33, 1 user, load average: 0.14, 0.23, 0.27
Tasks: 128 total, 9 running, 119 sleeping, 0 stopped, 0 zombie
Cpu(s): 12.6% user, 2.0% system, 0.0% nice, 85.4% idle
Mem: 1025820k total, 1006876k used, 18944k free, 201844k buffers
Swap: 2097136k total, 2288k used, 2094848k free, 635496k cached

Using tmpfs i rise down the cpu load.

I am recording all of he calls.

I haven´t got nothing special installed there.

slackware + scratch install.


I disabled the mix and compress cron entries. I only doit when the agents are not present.