Service deployed in priority resources phase

[joshuasimon-taulia]

Bug report

when i krane deploy an app for the first time, it times out because the service selects 0 pods. after kubectl deleteing the Google Config connector services.serviceusage.cnrm.cloud.google.com CRD, the deploy works.

Expected behavior: Service deploys after the priority resources phase

Actual behavior:

krane deploy na1prd na1prd --no-prune -f output/myapp/templates/
/Users/joshua.simon/.rvm/gems/ruby-head/gems/mime-types-3.3/lib/mime/types/logger.rb:30: warning: `_1' is reserved for numbered parameter; consider another name
/Users/joshua.simon/.rvm/gems/ruby-head/gems/mime-types-3.3/lib/mime/types/logger.rb:30: warning: `_2' is reserved for numbered parameter; consider another name
/Users/joshua.simon/.rvm/gems/ruby-head/gems/mime-types-3.3/lib/mime/types/logger.rb:30: warning: `_3' is reserved for numbered parameter; consider another name
[INFO][2020-07-10 10:32:56 -0700]
[INFO][2020-07-10 10:32:56 -0700]   ------------------------------------Phase 1: Initializing deploy------------------------------------
[INFO][2020-07-10 10:32:59 -0700]   All required parameters and files are present
[INFO][2020-07-10 10:32:59 -0700]   Discovering resources:
[INFO][2020-07-10 10:33:02 -0700]     - Deployment/na1prd-myapp
[INFO][2020-07-10 10:33:02 -0700]     - Ingress/na1prd-myapp
[INFO][2020-07-10 10:33:02 -0700]     - Service/na1prd-myapp
[INFO][2020-07-10 10:33:02 -0700]     - HorizontalPodAutoscaler/na1prd-myapp
[INFO][2020-07-10 10:33:03 -0700]
[INFO][2020-07-10 10:33:03 -0700]   ----------------------------Phase 2: Checking initial resource statuses-----------------------------
[INFO][2020-07-10 10:33:05 -0700]   Deployment/na1prd-myapp Not Found
[INFO][2020-07-10 10:33:05 -0700]   HorizontalPodAutoscaler/na1prd-myapp Not Found
[INFO][2020-07-10 10:33:05 -0700]   Ingress/na1prd-myapp Not Found
[INFO][2020-07-10 10:33:05 -0700]   Service/na1prd-myapp Not found
[INFO][2020-07-10 10:33:05 -0700]
[INFO][2020-07-10 10:33:05 -0700]   ------------------------------Phase 3: Predeploying priority resources------------------------------
[INFO][2020-07-10 10:33:07 -0700]   Deploying Service/na1prd-myapp (timeout: 420s)
[INFO][2020-07-10 10:33:39 -0700]   Still waiting for: Service/na1prd-myapp
[INFO][2020-07-10 10:34:09 -0700]   Still waiting for: Service/na1prd-myapp
[INFO][2020-07-10 10:34:39 -0700]   Still waiting for: Service/na1prd-myapp
[INFO][2020-07-10 10:35:09 -0700]   Still waiting for: Service/na1prd-myapp
[INFO][2020-07-10 10:35:42 -0700]   Still waiting for: Service/na1prd-myapp
[INFO][2020-07-10 10:38:11 -0700]
[INFO][2020-07-10 10:38:11 -0700]   -----------------------------------------Result: TIMED OUT------------------------------------------
[FATAL][2020-07-10 10:38:11 -0700]  Timed out waiting for 1 resource to deploy
[FATAL][2020-07-10 10:38:11 -0700]
[FATAL][2020-07-10 10:38:11 -0700]  Service/na1prd-myapp: GLOBAL WATCH TIMEOUT (300 seconds)
[FATAL][2020-07-10 10:38:11 -0700]  If you expected it to take longer than 300 seconds for your deploy to roll out, increase --global-timeout.
[FATAL][2020-07-10 10:38:11 -0700]
[FATAL][2020-07-10 10:38:11 -0700]    - Final status: Selects 0 pods
[FATAL][2020-07-10 10:38:11 -0700]    - Events: None found. Please check your usual logging service (e.g. Splunk).

Version(s) affected: krane 1.1.1 and krane 1.1.4

Steps to Reproduce

1. Install Google Config Connector CRDs
2. krane deploy this yaml

---
# Source: myapp/templates/service.yaml
apiVersion: v1
kind: Service
metadata:
  name: na1prd-myapp
  labels:
    app.kubernetes.io/name: myapp
    helm.sh/chart: myapp-0.1.0
    app.kubernetes.io/instance: na1prd
    app.kubernetes.io/version: "1.0"
    app.kubernetes.io/managed-by: Helm
spec:
  type: NodePort
  ports:
    - port: 80
      targetPort: http
      protocol: TCP
      name: http
  selector:
    app.kubernetes.io/name: myapp
    app.kubernetes.io/instance: na1prd

[dturn]

Thanks for the bug report. It looks like our logic to determine what is a priority resource only considers kind and not group so its confusing services.serviceusage.cnrm.cloud.google.com (which is a CRD and should be predeployed) with service.core which shouldn't.

[joshuasimon-taulia]

thank you!