hazelcast example is broken

[caesarxuchao]

I followed the instructions in the tutorial and got this output:

$kubectl get pods hazelcast
POD         IP           CONTAINER(S)   IMAGE(S)              HOST                                     LABELS           STATUS    CREATED         MESSAGE
hazelcast   10.244.0.5                                        kubernetes-minion-7jsw/130.211.122.180   name=hazelcast   Running   About an hour   
                         hazelcast      pires/hazelcast-k8s                                                             Running   9 seconds       last termination: exit code 1

Note the "last termination: exit code 1".

And here is the log.

It seems to be a problem of hazelcast rather than Kubernetes. @pires @nikhiljindal

[nikhiljindal]

@pires: Cant assign the issue to you, but I assume you are the right person to look at this?

[pires]

@nikhiljindal I will take a look, yes but this has been changed recently, even by the user who just opened this ticket. How did his PR got merged without testing, I wonder?

@caesarxuchao what version of Kubernetes is this?

[caesarxuchao]

It's 0.16.0.

In that PR, I updated the .yaml files to use the latest API version (v1beta3). I tested the original .yaml files, which are using v1beta1 API, and it causes the same problem.

@nikhiljindal, maybe revert that merge until @pires has a look?

[pires]

@caesarxuchao my only question is why merge something that wasn't working in the first place? No harm done. I will try with your changes and without, but most probably it's something related to how the apiserver replies now and the library I'm using.

[caesarxuchao]

@pires I will take caution next time. Thanks.

[pires]

PR is up, please review.

[caesarxuchao]

aha, endpoints contain a list of subsets...that's why my fix to cassandra is not working yet

[pires]

@caesarxuchao yeah, tried to discuss that with you over IRC, but you don't seem to have received my messages.