I have a fail-over cluster with two host servers and one connection broker. Each server is running Server 2016. Within the RDS deployment, there are multiple pools for individual workstation pools. The virtual machines in the pools are running Windows 10. We have a similar deployment for many sites across our company. They work pretty flawlessly most of the time. However, occasionally an issue arises where all users will receive the "requested session access is denied". Upon investigating, the services seem to be running, as expected. The virtual machines in the pools are running, and I can connect to them via Hyper-V Manager. There is just some disconnect that arises between the Connection Broker, the RDS Deployment, and the vms. The quick fix I've found is to restart the Connection Broker, which resolves the issues. I've tried restarting only the Remote Desktop Connection Broker service, but that didn't correct the problem.
In the event logs, at the time of the problem, I'm seeing:
RD Connection Broker could not find an assigned Personal Virtual Desktop for user XXXXX. HRESULT = 0x80070005
and
RD Connection Broker failed to process the connection request for user XXXXX. Load Balancing failed or Specified endpoint could not be found. Error: Access is denied.
Again, this is for ALL users across the deployment, and the issue is corrected when restarted the connection broker. I'm stumped. Has anyone seen this before or have any insight?