After x weeks of operation recordings are terminated prematurely due to timeout on passive, SPAN-based direct NIC capturing solutions with specific HW/NIC/Virtual env conditions.
In some conditions NIC/winpcap and OS clock is out of sync and over time (more weeks) the clock drift can accumulate and reach call timeout threshold.
- only impacts Passive SPAN-port based recording
- does NOT affect Media Collector source and remote capturing on Media Collector (most Lync recording deployments)
- the issue happens on rare combinations of specific platforms depending on NIC, NIC driver and virtualization environment (cases found in virtualized, but SPAN port based systems)
- it might happen several weeks after recorder service has been up and running (depends on hw)
- recording sessions terminated prematurely (before call ends) with end cause: call timeout (calls are only partially recorded)
- Set higher call inactivity/timeout threshold (def 180 sec -> 600 sec)
- Restart recorder service every weekend