Skip to content

Conversation

@travisghansen
Copy link
Contributor

Fixes #23

This attempts to address the issue by using a separate counter for VMs that are allocated to run.

It is possible to have more than 2 Pods submitted for the node (most easily if you simply manually set nodeName on the Pod spec). This addresses those scenarios.

I'm not familiar enough with the internals of the app to know if the decrement logic is appropriate. Most specifically:

  • I do not know when DeleteVirtualMachine is invoked and if that could be overly delayed. For example if the Pod is still in k8s but completed would DeleteVirtualMachine be invoked?
  • If vm.Start() fails in DeleteVirtualMachine invoked? Is it invoked immediately or after some other event triggers it?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

incorrect logic around MaxVirtualMachines

1 participant