Hi All
I am having quality issues, and am hoping for input/tips/advice on the setup of the zaptel conf files, and on how to diagnose the source of the problems.
Background
asterisk 1.2.11
zaptel 1.2.6
about 40 sip polycom 501’s
several fax machines on sipura modems
2 active iaxmodems running hylafax
AMP
box is 1.3ghz processor, 1gig ram, 160gig HD, 1u
heartbeat
one TE110P digium card, additional network card
I receive the last four digits on the t1, no caller id (flexgrow)
All systems work properly, the faxes come in flawlessly, but the sound and user experiences are an issue. There might be potentially ten active extensions at peak periods. The processor usage bounces around, generally low, but occasionally spiking to 80-90% or so, for brief moments. I have noticed additional unused asterisk processes showing up, but the problems persist even when I have eliminated those.
Yesterday and today the system seemed to get progressively worse, reloading it did not help, but a restart of the system set it straight, sufficient enough to finish the day, with some of the issues. Problems included one way audio when receiving a call (both sip to sip but mainly zap to sip and sip to zap), dropped calls, occasional echo, some volume issues, and then when things proceeded to worse…the time conditions would fail, sending the incoming calls to the incorrect destination, calls no longer rang to the operators etc etc. Basically the system could not be used sufficiently at that point.
Since a restart fixed the issues, temporarily, that leads me to the zaptel layer.
Original setup with the above problems
/etc/zaptel.com
span=1,1,0,esf,b8zs
e&m=1-24
/etc/asterisk/zapata.conf
signalling=em_w
I understood the flexgrow T1 to use wink start, and the above settings seemed to work. I added callprogress=yes to zapata.conf to help * answer certain lines that it could not detect the call being answered. That was an answer detect issue rather than an audio issue.
After today’s issues surfaced, and the testing ended, I spoke to the T1 provider, Verizon, Tampa, FL, and they outlined the provisioning they had on their data sheets as:
coding: eb8zs/sf
framing: d4
They could not say much more.
From this, I believe I may have sf signalling, but cannot figure out how to edit zaptel to make it work. From my research, this is a working (untested) configuration:
New Configuration
/etc/zaptel.conf
span=1,1,0,d4,b8zs
fxoks=1-24
/etc/asterisk/zapata.conf
signalling=fxo_ks
Can you offer any insight to what Eb8zs/sf means?
Studying the log file /var/log/messages and asterisks log file does not show me anything that seems out of the ordinary.
One question is, how to monitor the degrading system? Better, how to fully tune the system to be at its best for my next test (before the users use it) so as to overcome the recent failings?
I have discovered zttool and ztmonitor … are there any others I should use?
Any thoughts on how to test and tune prior to letting the users test it would be appreciated.
Thanks
Jonathan