I am Locutus of Borg

I’ve been diving head first into the k8s ecosystem over the last month (March 2017). This is the 2 or 3rd time “diving in” over the last 3 years and this time is the same as the last times. My experiences have been unpleasant .. here is a running list of the problems i’ve encountered.

  • Every time I want to view the dashboard I get “unauthorized” .. requires you run proxy in some cases

  • <pending> experience on spark - it always seems like external ip is sitting <pending> .. why is this
  • Are there any native load balancers or just GCE?
  • GCE gets more focus than native and other clouds.. i think this is unfortunate
  • The “Documentation” is all over the map.. old / new .. documention is either nonexistant or very confusing given the frequency of change
  • PVC->PV doesn’t seem to work with this “missing alpha thing” error

  • Pods stuck in ‘terminating’ status like this

  • Hitting CPU limits on the DNS container in k8s and getting a hundred containers built that need to be cleaned up

  • Adding credetionals for to .docker/config.json in minikube
  • Helm tiller node fails in helm or is overwelmed .. what to do? i ended up getting rid of minikube entirely and starting over
  • Container image ‘pull’s get stuck and minikube needed to be restarted .. happend on quay/stackanetes/mariadb image

  • A helm install fails after helm delete.. needed a –purge.. not obvious

  • What’s with randomly changing the order of output when you ask for multiple resources at once?

Granted, I’m a relative newbie. Thankfully there are plenty that have gone before me.. but clearly, there is work to do.

comments powered by Disqus