[Home]

Summary:ASTERISK-12701: Queues with Dynamic members like "Local/" can't get status of member and sometime get second call from queue
Reporter:Badalian Vyacheslav (slavon)Labels:
Date Opened:2008-09-09 06:07:11Date Closed:2011-06-07 14:07:23
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Applications/app_queue
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:Hello all.

Queues with Dynamic members like "Local/" can't get status of member and sometime it's get second call from queue.

Example:

have queue test
its have 4 members
1. Local/1001@agents (paused)
2. Local/1002@agents (free)
3. Local/1003@agents (In Use)
4. Local/1004@agents (In Use)

and have 3 queue callers.
1. Callerid 101
2. Callerid 102
3. Callerid 103

Strange what some time (above 30-60% of calls) free agent get Second call from queue (callerid 102).
Also in show queue - all member anytime is "Not In Use" (i use rewriten example from 1.4 version to mirgate to 1.6 without AgentLoginCallback).

In "Additional Information" exapmle of my agents content..

Thanks!

****** ADDITIONAL INFORMATION ******

context agents {
       _XXXX => {
               Set(device=${DB(agents_vs/${EXTEN})});

               if( ${GROUP_COUNT(${EXTEN}@agents)}=0){
                       Set(OUTBOUND_GROUP=${EXTEN}@agents);
                       Set(peer=${CUT(device,\/,2)});
                       Set(device_ip=${SIPPEER(${peer}:ip)});

                       Dial(${device},20,tg);
                       switch(${DIALSTATUS})
                       {
                          // some work with NO ANSWER and another states to autologoff and autopause and work with MySql
                       }
               } else {
                       Busy();
               }
       }
}
Comments:By: Doug (doug) 2008-09-09 11:45:53

We have this issue as well and have had it every now and then for ages. We have also had where the status stays as "in use" and no calls will be passed. We have to reload the queue app to reset the queue states.

slavon - are you using state_interface patch?

Do you see messages like:

[Sep  8 16:55:36] WARNING[13905] app_queue.c: The device state of this queue member, Local/1841@default-agent/n, is still 'Not in Use' when it probably should not be! Please check UPGRADE.txt for correct configuration settings. ?
By: Badalian Vyacheslav (slavon) 2008-09-10 00:46:12

DougUDI no, i did not use state_interface patch... yes, i see messages like this =)

I do not never see "in use" state in "show queues"... it allways Not In Use...  i detect state by perl script from queue_log... GROUP_COUNT work fine for detect and limit calls to "in use" agent.

In Use problem not critical for us but i think that it guilty for main problem. Somtime agent get Second call from queue! If hold time of first memeber is 10 minutes and second 3 minutes its very wrong for clients =(
By: Doug (doug) 2008-09-10 02:31:58

We have also seen problems with wait times on incoming calls and agents getting more than 1 call. This is a problem on high load systems as the queue will continuously hand calls to the agents even when they are on a call, this causes a large amount of traffic  and channel creation as well as very many entries in the CDR logs. We are currently changing to Direct agents to see if we have the same problem. I asumed at 1st this problem was with the state_ interface patch but seems i was wrong if you are having the same without using it.
By: Badalian Vyacheslav (slavon) 2008-09-10 03:56:35

nope... our agents not get more than 1 call... GROUP_COUNT work fine for this situaion... our free agent get second call from queue, but fist call (with most holdtime) was still in queue and not answered...
By: Doug (doug) 2008-09-15 02:31:12

Ok well I have now changed to Direct members and I still have the same problem. The system seemed to run fine for a few days and suddenly today has started causing a problem again. We are still seeing

[Sep 15 09:30:59] WARNING[9659]: app_queue.c:3069 try_calling: The device state of this queue member, SIP/1810, is still 'Not in Use' when it probably should not be! Please check UPGRADE.txt for correct configuration settings.

and the queue passes more that 1 call at a time to an agent thinking they are not in use.
By: Leif Madsen (lmadsen) 2008-09-15 12:53:07

I believe this is looking like a configuration issue, but I'd like more information to confirm that.

Please post relevant parts of your dialplan (extensions.conf), sip.conf, and queues.conf.

Do you have hints enabled in the dialplan? Do you have call-limit=1 (or more) enabled in sip.conf for your peers?
By: Leif Madsen (lmadsen) 2008-09-15 12:54:22

You might also find this works if you change:

queues.conf
-----------

member => Local/1001@agents

to...

member => Local/1001@agents/n


In addition to the things I mentioned in my last note. You should not be seeing that message, and if you are, it is probably part of the reason you are getting multiple calls.
By: Badalian Vyacheslav (slavon) 2008-09-16 00:38:15

blitzrage - sorry but bugtrack open by me =)

i post my dialplan upper.
call-limit=8

My main problem that agent some time get second call from queue...
By: Leif Madsen (lmadsen) 2008-09-16 12:13:31

I'm nearly convinced this is a configuration issue, and not a bug in Asterisk. Please utilize the available online help communities on the asterisk-users mailing lists, and the #asterisk IRC channel on the Freenode network at irc.freenode.net.

Once you have taken that route and determined it is a bug and are able to reproduce the issue with enough information, you may find a bug marshall on #asterisk-bugs to reopen this issue.

Thanks for using Asterisk!