Outbound call from Lync fails – TimeOut Issues

I was working on a strange issue at a customer regarding Enterprise Voice from Lync.
The issue:
– Some calls fails before call setup completes !!

Lync environemt was configured with standalone Mediation server, with direct SIP connection to Telecom provider.

The issue was reported to the Telecom provider, and they responded quickly with a knowledge of this issue.
The response was:

The Lync Mediation Server is sending a CANCEL on call setup, after a very short time.
After some update from Microsoft, Lync has become unpacient, and if the remote party hasn’t reponded with more that “100 Trying” during 10 sec the Mediation Server sends a CANCEL !!
This timer was earlier 30 – 40 seconds, but is now only 10 !!
The remote party can’t respond with more that “100 Trying”, until they have received anyting from the Called Party.

The best part is that the Telecom provider has been in contact with Microsoft regaring this, and they got a FIX 🙂

The fix:

Configuring Parameters

Some of the above timeouts can be configured. The file which has the configurable parameters is ‘OutboundRouting.exe.config”  Use caution when changing these values, as a rule of thumb try not to increase or decrease the value by more than 25% of its original value.

From OutboundRouting.exe.config

<configuration>

    <appSettings>

      <add key=”FailOverTimeout” value=”10000″/> – The problem line 🙂

      <add key=”MinGwWaitingTime” value=”1″/>

      <add key=”MaxGwWaitingTime” value=”20″/>

      <add key=”FailuresForGatewayDown” value=”10″/>

      <add key=”FailuresForGatewayLessPreferred” value=”25″/>

      <!– Valid values are between 5 and 600 –>

      <add key=”HealthMonitoringInterval” value=”300″/>

      <!– Valid values are between 60 and 3600 –>

      <add key=”GatewayStateReportingInterval” value=”1800″ />

  </appSettings>

</configuration>

The FailOverTimeout should be increased to the desired time limit.
The file is found under C:\Program Files\Microsoft\Lync Server 2010\Server\Core on the FrontEnd Server.

Changing the value from 10000 (10 sec) to 20000 (20 sec) solved the issue.

After changing this value, it’s recommended to reboot the server.