CCE 2.1.0 - Draining Calls

The Cloud Connector Edition system for Skype for Business online is a very impressive collection of scripts and glue and luck that has the ability to build a complete SfB voice infrastructure and patch it automatically. One of the parts of this is not as good as it could be, read on if you get call drops when updating :-)

Environment is 2x CCE hosts in the same site. For the test below I first put the CCE2 into maintenance and make calls. Therefore forcing all calls through CCE1.

I make two calls:

Ensure that both calls are up and running and show on the Sonus:


Logging into the Mediation Server on the CCE1 you can see the calls are running through it:



Take CCE2 out of maintenance so that both CCE's take service calls (both calls are still running on CCE1).

On CCE1 run the command Enter-CcUpdate on an elevated PowerShell session

Based on the documentation (https://technet.microsoft.com/en-us/library/mt492520.aspx) I would expect this to drain the mediation server, waiting for the two nailed up calls to complete and to then gracefully stop the services.

Instead I get this:



Services are stopped and both calls are dropped. Bit of a fail when the documentation says:

"The appliance is “drained”—that is, all existing calls will complete, but new calls are rejected."

and

"The Enter-CcUpdate cmdlet will ensure that all running calls on a Cloud Connector appliance will complete, but the appliance will reject any new calls, which are transferred to other production appliances. This cmdlet enables you to update an appliance without affecting end users calls." (my emphasis!)

(Bonus points for the spelling of "Drainning" and "Forceing")

Its now logged as a ticket with Microsoft and I'll update as and when I have a resolution.


Update: 22nd January 2018
Confirmed as a bug and passed to Product Group to address. Workaround is to connect to Mediation Server and perform a Stop-CSWindowsService -Graceful command





No comments:

Post a Comment