Adding static agents to queues breaks queues on startup

you'll have to be a little more specific (I've got my trigger finger on worksforme...) I did a restart with static agents, all is fine. I then added a new static agent, hit the reload bar, then did a restart of asterisk, all is still fine - no problems. So can you please provide exact repro steps as well as providing specific Asterisk version number...

Don't restart asterisk. exit asterisk, service asterisk stop and then service asterisk start

Then look do a show queues this is where it's doing it for me.

Then if I simply type reload everything works properly.

If you want, I can upload a freepbx backup with the config, I can duplicate this over and over on any of our customers systems, or I can let you ssh into a dev system so you can see for yourself.

Again, specific version number.

Next - if you are doing 'service asterisk stop/start' you are technically not using the supported method of starting/stopping asterisk. However, with that said, the amportal start/stop simply does a graceful stop and start of Asterisk. What you describe sounds like an issue with Asterisk somewhere so more Asterisk specifics will be needed.

Also, please test the following and report back along with the version information. Please try your issue by removing: "persistentmembers=yes" from queues_general_additional.conf and then see if the problem continues. That is about the only recent change to queues (unless you are using autofill or queue weights which were also added but configurable).

In addition, you should try and load a non trixbox system to see if the issue can be reproduced, because if it turns out to be on trixbox only and something with Asterisk, it's unlikely anyone on the Asterisk team will be willing to look at it since they are RPM binary builds that are non-standard (I think) which means too many unknowns.

For what it's worth though (running 1.4.21.1) I did the following, where I already have static agents in 2 queues, and one of the queus has 2 dynamic agents currently logged in:

amportal stop
amportal start
CLI> queue show
(all looks fine)
amportal stop
service asterisk start
service asterisk stop
service asterisk start
CLI> queue show
(all looks fine)
service asterisk stop
service asterisk start
CLI> queue show
(all looks fine)
service asterisk stop
amportal start
CLI> queue show
(all looks fine)

So ... the worksforme trigger finger is being restrained, since I'm curious what may be going on with Asterisk...

I'm going to try pbxinaflash 1.2 and get it updated to similar asterisk versions and see how that goes.

I loaded trixbox 2.6.0.7 elastix and pbxinaflash

I never updated anything

on all three all I did was the following: 1. Added queues module 2. created ext 100 with name reception, and enabled voicemail 3. created queue 150 with name reception 4. applied changes. 5. amportal stop 6. amportal start asterisk -r show queues

and all 3 of them showed invalid queues.

I Am seeing this behaviour on asterisk 1.4.18

That being said, as soon as a call hits the queue, If the agent is available(as in Valid Extension and Logged in), Their status changes to Not in use from invalid.

So even though it probably isn't correct, it doesn't break anything.

It does break it.

If the customers server looses power and comes back up, the calls will not enter the queue and go to the fail over destination.

It does not work properly until we log into the system and simply reload the configuration.

we have our customers setup that calls come into the queue first and then hit the ivr, so it's very important that this works. Customers that dynamically login have no problems but some customers want the receptions phones in their statically.

Also on our systems it *never* changes from invalid to anything else.

I am unable to replicate your fault on 1.4.18 or 1.4.21.2 with Freepbx versions 2.5.0rc1 and 2.4.1 respectively.

On 1.4.18 I can see the agent status as invalid, but when the first call comes in, it updates the agent status.

On 1.4.21.2 the agent comes up as not in use. These are both Roll your own installs on Centos5

kspare, I'll monitor if there is more data and we can reopen this if it can be reproduced, but it looks from other feedback that all is working properly.

p.s. you never tried any of the tests I requested wrt to removing persistentmembers to see if that had any effect. It really helps when reporting bugs like this, to try the suggested changes from the devs as well as provide the information such as Asterisk version, so that we can try to get to the bottom of these types of issues. Thanks for what you have done so far though.