search

cncf-infra / coder-templates

3 stars 4 forks source link

Add support for suspending KubeVirt VMs instead of deletion during coder stop #11

Open hh opened 1 year ago

hh commented 1 year ago

This will likely require research using count = data.coder_workspace.me.start_count set appropriately on resources that need to be spun down vs persisting through shutdown.

Since most of our kubernetes resources are created dynamically via cluster-api, we will need to leave them 'up' and only run what is necessary to suspend the created VMs.

I would suggest figuring out how to suspend / restore the created KubeVirt VMs without updating the template first. Then work back and add support for the appropriate terraform resource to do the same.

BobyMCbobs commented 1 year ago

I thought about setting the TalosControlPlane.spec.replicas field to 0 but it appears to not be allowed

   - lastTransitionTime: "2022-12-07T21:29:04Z"
     message: Cannot scale down control plane nodes to 0%!!(MISSING)(EXTRA *int32=0xc0017106c0,
       int=1)
hh commented 1 year ago

I think it’s probably fine to shutdown the VM and start the VM itself.

Or if we want to be really fancy : https://github.com/kubevirt/kubevirt/issues/4573

On Wed, 7 Dec 2022 at 5:15 PM, Caleb Woodbine @.***> wrote:

I thought about setting the TalosControlPlane.spec.replicas field to 0 but it appears to not be allowed

  • lastTransitionTime: "2022-12-07T21:29:04Z" message: Cannot scale down control plane nodes to 0%!!(MISSING)(EXTRA *int32=0xc0017106c0, int=1)

— Reply to this email directly, view it on GitHub https://github.com/cncf-infra/coder-templates/issues/11#issuecomment-1341724408, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAHUY6ORS7H2TPJEF2NGBLWMELCZANCNFSM6AAAAAASWX23BQ . You are receiving this because you authored the thread.Message ID: @.***>