pod-culler upgrade - ServiceAccount required before initialized?

[consideRatio]

I toss this issue out before I go to bed, got any suggestions I'm happy to hear about them. I'll spend more time figuring things out later.

---

I just attempted to update to a more recent release of this repo, but got the following error.

Error: UPGRADE FAILED: no ServiceAccount with the name "pod-culler" found

It makes little sense to me, I can see the rbac.yaml definition within pod-culler template directory. I think my deployment is conditioned out by my cull.maxAge being 0.

# my values.yaml
rbac:
  enabled: true

cull:
  enabled: true
  users: false
  timeout: 3600
  every: 600
  maxAge: 0

proxy:
  https:
    enabled: true
    type: letsencrypt
    letsencrypt:
      contactEmail: 'myemail@host.com'
    hosts: [mydomain.se]

zero-to-jupyterhub-k8s/jupyterhub/templates/pod-culler/deployment.yaml

Lines 1 to 5 in d7d7a58

	{{ if and .Values.cull.enabled .Values.cull.maxAge }}
	apiVersion: extensions/v1beta1
	kind: Deployment
	metadata:
	name: pod-culler

zero-to-jupyterhub-k8s/jupyterhub/templates/pod-culler/rbac.yaml

Lines 1 to 5 in d7d7a58

	{{ if .Values.rbac.enabled -}}
	kind: Role
	apiVersion: rbac.authorization.k8s.io/v1beta1
	metadata:
	name: pod-culler

[consideRatio]

OH! perhaps we need to move the ServiceAccount from the bottom of the rbac.yaml to the top?

[consideRatio]

Notes

• cull.maxAge: 0, but the conditional {{ if and .Values.cull.enabled .Values.cull.maxAge }} seem to be evaluated to true (ping @choldgraf --- something of relevance? I recall you had issues with the culler.)
• I had to delete a serviceaccount before it could be recreated again to get rid of the "ServiceAccount not found" trouble.
• Gaaaah whats going on with the "no .... of the name .... found" ?!?!??!!? I figure I need to cleanup helms release number history or similar...

[consideRatio]

Resolved --- By using helm 2.8.2 with a clean slate. To learn more see #620