Manual dial issues in vicidial_agent_log

[jjohnson78]

Asterisk version: 1.4.38
Vicidial VERSION: 2.6-365a
BUILD: 120420-1620

I'm having an issue with the vicidial_agent_log table - it has apparently been going on for a while. We have several campaigns that are using the INBOUND_MAN dial method, and on one a rep placed a call and dispoed it as a SALE.

The call was logged in the vicidial_log table properly, as below:

uniqueid: 1345489108.411
lead_id: 54510494
list_id: 12000
campaign_id: SKINABDN
call_date: 2012-08-20 14:58:42
start_epoch: 1345489122
end_epoch: 1345489632
length_in_sec: 510
status: SALE
phone_code: 1
phone_number: 519*******
user: 37327
comments: MANUAL
processed: N
user_group: VICI01
term_reason: AGENT
alt_dial: MAIN

However, when the rep placed this call, the record that was put into the vicidial_agent_log table was not updated - this was the corresponding record in the vicidial_agent_log table:

agent_log_id: 122335828
user: 37327
server_ip: 10.10.12.11
event_time: 2012-08-20 14:58:27
lead_id: 54510494
campaign_id: SKINABDN
pause_epoch: 1345489107
pause_sec: 1
wait_epoch: 1345489108
wait_sec: 1
talk_epoch: NULL
talk_sec: 0
dispo_epoch: NULL
dispo_sec: 0
status: NULL
user_group: VICI01
comments: MANUAL
sub_status: NULL
dead_epoch: NULL
dead_sec: 0
processed: N
uniqueid: (blank)


I assume the reason it wasn't updated is because the uniqueid was not part of the record that was inserted, but I have no idea why that would be happening. Also, the next call the agent placed had all 510 seconds from that call added to each of the four categories of pause, wait, talk and dispo seconds:

agent_log_id: 122335829
user: 37327
server_ip: 10.10.12.11
event_time: 2012-08-20 14:58:28
lead_id: 54510952
campaign_id: SKINABDN
pause_epoch: 1345489108
pause_sec: 543
wait_epoch: 1345489651
wait_sec: 557
talk_epoch: 1345489122
talk_sec: 510
dispo_epoch: 1345489650
dispo_sec: 540
status: NQ
user_group: VICI01
comments: MANUAL
sub_status: NXDIAL
dead_epoch: NULL
dead_sec: 0
processed: N
uniqueid: (blank)

Again, no uniqueid was logged, and the corresponding record for this call in vicidial_log shows the call was 0 seconds. What is also puzzling about this is that the call_date in the vicidial_log table is way off from the event_time in the agent_log table on both calls, but we are not getting time sync errors on the agent's interface, nor are the times on any of the servers (web, db, dialers) off from each other by more than a fraction of a second. Any ideas what could be causing this? This is causing the agent performance detail to be inaccurate as well, of course, as this is happening to several different reps in different INBOUND_MAN campaigns. Thanks!

[williamconley]

Post your installation method and a CLI output from an occurrence. You'll probably have to run a few controlled tests to get it to happen. But if you simulate the call path it should not be too hard.

Please don't post 3000 lines of unrelated code. Run the test with no other traffic and only post the output from this call.

You may also need to post agi output from the appropriate log in /var/log/astguiclient