Nova multi-cell adoption requires different renaming technics for cells databases during importing it

[bogdando]

• Document deployment of dev setup for adoption of a multi-stack topology for Nova cells v2.

  • Gotchas: connecting an extra standalone compute host to the central standalone host is problematic as it doesn't fit the supported multi-cell v2 topologies in OSP. No dev envs shortcuts possible? If that would become supported at least for dev envs, the approximate scenario might look like:

    • Create the edpm-compute-1 and edpm-compute-2 virtual machinee for Nova compute cells.

    make edpm_compute EDPM_COMPUTE_SUFFIX=1
    make edpm_compute_repos EDPM_COMPUTE_SUFFIX=1
    

    • Omit the edpm_deploy make target to not making it managed from the control
      plane running on OCP. Instead, deploy it as a 2nd TripleO standalone Heat stack,
      with an extra OSP compute:

    make standalone EDPM_COMPUTE_SUFFIX=1 EDPM_COMPUTE_CEPH_ENABLED=false
    (requires intall_yaml to support exporting stack data to use it with consequent stacks additions 
    https://github.com/openstack-k8s-operators/install_yamls/pull/568)
    

    • Ssh into deployed standalone host:

    ssh -i ~/install_yamls/out/edpm/ansibleee-ssh-key-id_rsa [email protected]
    

    • Discover the deployed compute host keys (a passwordless access is required by
      the Nova Live-migration feature):

    ssh-copy-id -i ~/.ssh/id_rsa [email protected]
    

    • Discover the remote compute node from the central controller node ...
• Adjust the databases importing/renaming script for real multi-stack cases:
  • nova_cell1-> nova_cell2, ..., nova_cellN -> nova_cellN+1, nova-> nova_cell1.
  • if there were computes in the 'default' (int terms of tripleo) cell, then it needs to be imported and adopted like a normal cell,
    and renamed like 'default' -> cell<latest+1>. For a single stack scenario, that is cell1. And the command that renames the default cell name should become update nova_api.cell_mappings set name="cell<latest+1>" where name="default";'
  • For that, add a second map in the DB importing script, and loop cell_name_map["default"] = "nova_cell$x".