While I have not had the same call volume previous posters describe, I have a large number of queues (about 65 right now), although each queue only has about 8 static agents. We have had this issue happen twice now in the past month or two.
We are also (coincidence or otherwise) running an HP server, a DL380 G5 with 2GB RAM and dual CPUs. CPU usage is rarely an issue, and call quality is normally very good, so I don’t think we have a resource issue. We are running Trixbox 2.2 CE, upgraded with FreePBX 2.5.x, so we still have Asterisk 1.2 under the hood.
Today was the second time this happened. There were 3 calls in progress and two more calls in queue. All were in different queues. All queues use ringall, 20 seconds, with 5 second wait, no timeouts, periodic announcements every 1 minute 30 seconds, random hold music.
The two calls on hold ended up being there for almost 7 minutes by the time two of the other calls wrapped up. By almost 10 minutes in queue, they still had not rung the available agents, although those agents could call each other, so they weren’t DND.
After the third call wrapped up, I restarted asterisk, and calls started ringing through again.
These calls were coming in via an IAX2 trunk from our client, then routed to the agents’ X-Lite SIP softphones.
I’ve read through this thread, but when the problem was occurring I did not check CPU usage, and would not have known to look for lots of dialparties.agi processes. I did notice in the /var/log/asterisk/full a huge number of lines (63927) of
Of course, the timestamps were different, as well as the DEBUG numbers and the ending number on the trunk…
Since the problem seemed very similar to mine, I checked, and our queues_additional.conf is using “Local” to identify the members. But it also looks like the recent discovery of how the manager call is made may resolve the issue, but will either of these work on Asterisk 1.2? (We are investigating our upgrade path, but this is what we have for now) Or, do we have a different issue altogether? (If so, I’ll start a different thread and apologize for disturbing this one.)
TIA
DionV
(For historical note: In the past, we have had a total of 4 queues with 20 calls in progress and another 15 in queue, all in one queue, with a ringall strategy, with little degradation of service, running on a PIII-700MHz dual CPU system, using Trixbox 2.2 CE stock. All calls were coming in from SIP trunks. This happened from 8am to 9pm Monday to Friday for about 3 weeks. Never had an issue like this. 1) Testimony to the power of the software! 2) Don’t think it’s a lack of hardware resources.)