Stuck in retry() loop for elastickube-server

[Justin Garrison]

I haven't quite tracked down the logic yet but my deployment on kubernetes 1.2.4 is stuck "Setting up elastickube-server"

From what I can tell with a quick read through on the deploy script I'm stuck in retry() and it doesn't appear the counter is incrementing (it's been stuck spinning for ~10 minutes).

If I open a new shell and just run a simple test for the rc I can see it in the output

if ! kubectl --namespace=kube-system get rc elastickube-server; then echo "not found"; fi

Something in the retry loop isn't working for me. I'm going to try to debug a little bit more but thought I'd open the issue anyway to let you know.

For more info I'm running on-prem.

After stopping and running the script a 2nd time it made it through all the steps no problem. But there still remains the initial bootstrapping issue.

[Arnaud]

@rothgar thanks for trying elastickube.

I have seen a similar issue where the script does not detect when the RC is ready on the initial bootstrapping.

Could run the command kubectl --namespace=kube-system describe rc elastickube-server and paste the output here?

In the meantime, I am going to try to reproduce locally.

[Justin Garrison]

I'm not 100% positive this is unchanged from a default install (I probably changed some things when I was testing) but here's the output.

kubectl --namespace=kube-system describe rc elastickube-server
Name:           elastickube-server
Namespace:      kube-system
Image(s):       elasticbox/elastickube-api:latest,elasticbox/elastickube-charts:latest,elasticbox/elastickube-nginx:latest,elasticbox/elastickube-diagnostics:latest
Selector:       name=elastickube-server
Labels:         name=elastickube-server
Replicas:       1 current / 1 desired
Pods Status:    1 Running / 0 Waiting / 0 Succeeded / 0 Failed
Volumes:
  elastickube-charts:
    Type:       HostPath (bare host directory volume)
    Path:       /var/elastickube/charts
  elastickube-run:
    Type:       HostPath (bare host directory volume)
    Path:       /var/run/elastickube
No events.

[Alberto Arias Maestro]

@rothgar The default timeout is 10 minutes. What we think it might have happened is that you aborted it right before the timeout was reached. The script is idempotent, so the second time you ran it found the RC already running.

We tried setting the timeout to something very small, and we still could not repro. Please let us know if you find anything.