ManageIQ
[WIP] [POC] Remove multiple server support from the orchestrator process.
This PR is a PoC to show how we don't need as_each_server at the evm_server level. It probably does not work as is. This is a partial revert of https://github.com/ManageIQ/manageiq/pull/19734
The idea here is that the orchestrator should only be responsible for its own "server". In appliances this is done automatically because we only watch one server anwyay, and we presume the rest of the server records are being watched by their orchestrators. In podified, this was changed to watch every "server", however all of the servers are fake records that are only in place to deal with the zone -> miq_server relationship which we need for zone support.
Instead, what we can do is just watch the one "server" record we care about, and then change the way we monitor other servers in server_monitor.rb. What's missing from this PR is that the server_monitor.rb code needs to change such that we can have a subclass for Kubernetes that effectively does nothing for the other "fake" servers (i.e. we don't care about heartbeats), but still allow for role management to support the zones concept of placement. @agrare had a good idea for a MiqServer::PlatformManagement, which I think aligns with this.
Built on #21532, so only look at the latest commit: https://github.com/Fryguy/manageiq/compare/Fryguy:remove_as_each_server~...Fryguy:remove_as_each_server?expand=1
cc @agrare @jrafanie
This pull request has been automatically closed because it has not been updated for at least 3 months.
Feel free to reopen this pull request if these changes are still valid.
Thank you for all your contributions! More information about the ManageIQ triage process can be found in the triage process documentation.