Queuemetrics says agents are logged out but they are not...

[rgrignon]

Here is the scenario:

1. Agents are all logged in via the Agent Page. Their status is accurate...
2. Asterisk Crashes or gets restarted
3. The following log is written to queue_log (and sent via qloaderd):
                                   1267176428|NONE|NONE|NONE|QUEUESTART|
 
4. At this point all of the agents status shows as "Logged Out" Even though they are still logged in...
5. I noticed that you are able to ignore QUEUESTART commands but changing this setting does not have any effect on my situation.

#Ignore QUEUESTART statements.
default.ignoreQueueStarts=false

I have tried
default.ignoreQueueStarts=true

and
default.ignoreQueueStarts=


You can replicate this issue by doing the following:

1. Log in via agent page
2. issue a "restart now" from the asterisk CLI
3. Refresh your agent page

[QueueMetrics]

This looks like a bug - which version of QM are you running? do you ise ADDFMMEBER or AGENTCALLBACKLOGIN to log on agents?
The behavior is expected to be configurable....

[rgrignon]

QueueMetrics - 1.5.5

I use AddQueueMember

[QueueMetrics]

Thanks - we are trying to reproduce the issue.
Will let you know how it goes.

[marcos]

Hello.

I've replicated the problem. I've inserted a bug in our internal bug tracker. You can refer it to #1002 for any other information.

Thank you and regards,
Marco Signorini.

[marcos]

Hi,
In my test setup I can confirm QueueMetrics is working fine if you set:

default.ignoreQueueStarts=true
default.useRawAgentSessions=true

Then issue a QueueMetrics restart.
Tell me if you have problems after these configuration modifications.

[rgrignon]

Still No good from my side and I even upgraded to the latest version...

Also, its not a "queuemetrics restart" that breaks it. It is a  "restart now" from the CLI (simulate a crash)

[fredsi]

Hi,

I have exactly the same problem. My supervisor page shows no agents logged on, but in CLI, I can see there are some agents logged on. I have to ask the agent do a remove member and add member again in order to show them in the supervisor page and their own page.

I have tried to change the configuration as suggested in previous post, but no luck. I don't see further posts. Is this problem fixed yet? BTW, I just installed QM recently, so it's a new version 1.6.0.1, and I'm using Elastix 1.5.2.

Another question, to add a single member, what level is it? Can I change that level? Thanks.

[QueueMetrics]

Still No good from my side and I even upgraded to the latest version...

Also, its not a "queuemetrics restart" that breaks it. It is a  "restart now" from the CLI (simulate a crash)

This is interesting - what do you get on the queue_log? can you simulate a case for us?

[fredsi]

thanks for the reply.

here is the situation: last week, i was testing the system, I put a sip phone in DND (do not disturb), and call it, my asterisk got restart, and i saw segfault in the log file, all the agents in QM show logged off, but in CLI, they still logged in. Today, i tried to do the same thing, still got same segfault, and asterisk restarted, but the QM status didn't changed, weird -- sorry can't duplicate the issue now. queue-log showed bunch of add and remove member:

1272049998|1272049998.45|101|Local/4071@from-internal|REMOVEMEMBER|
1272050001|1272050001.47|106|Local/4071@from-internal|REMOVEMEMBER|
1272050004|1272050004.49|103|Local/4071@from-internal|REMOVEMEMBER|
1272050007|1272050007.51|104|Local/4071@from-internal|REMOVEMEMBER|
1272050023|1272050023.55|101|Local/4071@from-internal|ADDMEMBER|
1272050026|1272050026.57|106|Local/4071@from-internal|ADDMEMBER|
1272050029|1272050029.59|103|Local/4071@from-internal|ADDMEMBER|
1272050032|1272050032.61|104|Local/4071@from-internal|ADDMEMBER|
1272050272|1272050272.71|101|Local/4053@from-internal|REMOVEMEMBER|
1272050275|1272050275.73|102|Local/4053@from-internal|REMOVEMEMBER|
1272050294|1272050294.77|101|Local/4053@from-internal|ADDMEMBER|
1272050297|1272050297.79|102|Local/4053@from-internal|ADDMEMBER|

now i still see couple agents logged on in CLI, but didn't show in QM real-time page, i believe they didn't log off yesterday. is there way to log them off everynight automatically?

thank you.

[QueueMetrics]

Basically the problem here is that the system dies without writing anything.
In case this happens, tell your agents to log off and on again.