Closed
Description
We recently moved to gci and using cloud-init instead of salt. This reverted our recent kube-dns rewrite by continuing to deploy the skydns rc, as is evident from this run: http://kubekins.dls.corp.google.com/job/kubernetes-e2e-gce/17559/consoleFull
https://pantheon.corp.google.com/storage/browser/kubernetes-jenkins/logs/kubernetes-e2e-gce/17559/
19:22:34 May 24 19:22:29.198: INFO: kube-dns-v11-m8wnv started at 2016-05-24T19:15:18-07:00 (0+4 container statuses recorded)
19:22:34 May 24 19:22:29.198: INFO: Container etcd ready: true, restart count 0
19:22:34 May 24 19:22:29.198: INFO: Container healthz ready: true, restart count 0
19:22:34 May 24 19:22:29.198: INFO: Container kube2sky ready: true, restart count 0
19:22:34 May 24 19:22:29.198: INFO: Container skydns ready: true, restart count 0
@wonderfly assuming #26197 flipped the switch
As a side note, we should surface cluster addon version somewhere in e2e-up output, I had to find a failed run and lookup the dumped pods to prove this point.
Activity