This page lists known issues for the Puppet orchestrator and orchestration services in Puppet Enterprise.
In this version of PE, the PCP broker uses immediate message delivery. If an agent node is disconnected, PXP will send 2 status requests in a row, each time immediately receiving a response that the agent is not connected, leaving no time for the pxp-agent on the agent node to restart and reconnect. This will be fixed in the next major version.
If you need to stop an orchestration job running on a Windows agent, you can use the /command/stop API endpoint.
When you use the Puppet orchestrator
puppet job command with the
environment flag, Puppet agent runs participating in the job should request the specified environment. A known issue exists where these agent runs may request a catalog for a different environment than the one specified in the job.
To workaround this issue, we recommend that you use the node classifier in the PE console to add agents to the environment that will be participating in Orchestration jobs. Alternatively, you could assign the agent’s environment in its configuration file.