table vicidial_closer_log value length_in_sec

[marzo]

Hello,

I have a cluster of Vicibox:
1. telephony server: Vicibox_v.6.0.x86_64-6.0.3 from .iso | Vicidial 2.12-505a Build150806-134 | Asterisk 1.8.32.3-vici | Sangoma A104E card | No Extra Software After Installation | ProLiant DL360 G7 | Intel(R) Xeon(R) CPU X5650 @ 2.67GHz
2. database server: Vicibox_v.6.0.x86_64-6.0.3 from .iso | Vicidial 2.12-505a Build150806-134 | No Extra Software After Installation | 440BX Desktop Reference Platform | Intel(R) Xeon(R) CPU E5-2680 0 @ 2.70GHz | VMware ESXi 6.0
3. web server: Vicibox_v.6.0.x86_64-6.0.3 from .iso | No Extra Software After Installation | 440BX Desktop Reference Platform | Intel(R) Xeon(R) CPU E5-2680 0 @ 2.70GHz | VMware ESXi 6.0

Just for some calls I got values in the vicidial_closer_log table where the value of length_in_sec is different than end_epoch minus start_epoch. By example:

start_epoch end_epoch length_in_sec status iniqueid
----------------- ---------------- ------------------- -------- -----------
1447076314 1447076602 281 XFER 1447076291.2378
1447077115 1447077379 198 DCMX 1447077099.2895
1447077376 1447077968 583 DCMX 1447077359.3114
1447079407 1447079789 368 XFER 1447079391.4527

The computers of the agents have his clocks synchronized with the telephony server.
The computers of the agents are locate in the same LAN but in different VLAN of the telephony server.
I have not seen a high latency between the computers of the agents and the telephony server.

Any idea about it ?

For all the calls the value of length_in_sec must be equal to end_epoch minus start_epoch ?

Regards