VxRack NeutrinoVxRack Neutrino Software
While testing Node Transfer In Neutrino 1.1, we discovered two major issues:
Issue # 1- After removing Cloud Compute service from a compute node, the unallocated node remains in the 'cc_compute' group.
Hence, when we transfer a Platform node to this unallocated node, the transfer playbook creates a 'cc_nova_compute' container on this new Platform Node.
And, it becomes part of the Hypervisors list.
Issue # 2 - As a consequence of the situation above, an instance can be hosted on the hypervisor running on the new Platform node.
However, when we transfer the current Platform node, the instances remain attached to the former Platform Node that become unallocated.
Also, this unallocated node (former Platform node) sill shows in the hypervisors list. But in a down / disabled state.
This puts the system in an incoherent state and the instances are not reachable anymore.
affected version 1.1
fixed in 22.214.171.124