Problem

The Agent disconnects from the Execution Server prior to or during task execution and generates the following error message in the Windows Application Event Log of the Agent machine:

HandleChannelDisconnection
System.ServiceModel.CommunicationObjectFaultedException: The communication object, System.ServiceModel.Channels.ServiceChannel, cannot be used for communication because it is in the Faulted state.

Note: For the latest on the fix, see the Update section below.

Affected Versions

The following versions are affected:

  • Automate BPA Server 10.5 and higher
  • Automate Enterprise 11.0 and higher

Solution

This issue will be fixed in the next maintenance release. The agent connectivity has been rearchitected from the ground up.

The HandleChannelDisconnect issue appeared with the release of v10.5 and is present in subsequent product releases as well. Our research has uncovered multiple root causes that have resulted in the same error message and agent disconnect, and many of these variants have been addressed with subsequent releases of v11. During this time, our development team has been working on rewriting the agent connectivity from TCP to HTTP to address the remaining variants.

If you are experiencing the HandleChannelDisconnect issue, please update to the latest version of v11 to see if the variant you are experiencing has been addressed. If the issue is still present after upgrading to v11, please apply the workarounds (see below) and await the next maintenance release. Customers with existing HandleChannelDisconnect issues will be notified when the maintenance release is available. We apologize for the inconvenience and delay this causes.

Update 1/10/2020

A majority of the customers participating in the controlled rollout are reporting success with the new code. The remaining issues reported by customers involved are under review. Upon completion of the testing cycle, development will start to build and conduct final testing on the GA release.

We are tracking towards an early February GA release. Another update to this bulletin will be provided by January 24th.

Update 12/26/2019

Internal development and testing has been completed. To insure the integrity of this fix, we're providing a controlled rollout instead of a full general audience release. We're currently performing the second round of the controlled rollout with designated customers. The initial results of customer testing have been positive.

We will provide another update to this bulletin by January 10, 2020.

Workarounds

Workaround #1

Configure the workflow tasks to run on each agent until it succeeds:

  1. Create an agent group and assign all agents to the agent group
  2. Reference all tasks within the workflow to run on the agent group
  3. Set the agent group distribution to “run on each agent in order until a successful result”.

Workaround #2 (Requires latest version of Automate Enterprise v11)

Configure the workflow items to retry when an error occurs:

  1. Go to Options > Server Default Properties.
  2. Set the “On Workflow Item Error” to “retry then continue”.

 


Still have questions? We can help. Submit a case to Technical Support.

Last Modified On: January 11, 2020