We create a temporary vBMC instance on the hypervisor to deploy OCP cluster via devscripts.

This instance gets removed once the OCP cluster is deployed.

It seems the OCP cluster information are pushed to the deployed metal3 ironic service, on the OCP cluster, and then ironic checks on the node status periodically. This, of course, fails, since vBMC is down at that time.

There's no quick workaround though:

• we can't just keep the vBMC as-is, since the VM name change, and it won't be able to find the VM, returning a bad status
• we therefore have to remove the deployment data, then inject the new ones with the new name, but same ports

While it's not creating any major issue, ppl are coming to us asking about those errors in ironic, and it might be slightly unsettling when they are trying to debug actual BM node provisioning...

Another solution would be to ask if, really, the OCP node data are to be passed down to the deployed ironic (imho it shouldn't, to avoid accident)...