[kubernetes-users] Running Calico on GKE

Hi guys, I would like to know if its possible to run Calico on *GKE* to enable fine-grained network policies ? If not, is it something planned ? Thanks for any advice ! -- You received this message because you are subscribed to the Google Groups "Kubernetes u

Re: [kubernetes-users] How to caclulate the resource usage in a particular namespace?

Thanks a lot, I will have a try. On Monday, November 28, 2016 at 11:02:42 AM UTC+8, David Oppenheimer wrote: > You can do a watch on the pods to be notified when they start and stop. This > tutorial appears to show you how to do it (I skimmed it and it looks about > right). > > > As an alterna

Re: [kubernetes-users] How to caclulate the resource usage in a particular namespace?

You can do a watch on the pods to be notified when they start and stop. This tutorial appears to show you how to do it (I skimmed it and it looks about right). As an alternative, the Kubernetes quota

[kubernetes-users] Too many ways to install, which one to pick?

I am trying to install k8s on AWS. There are three different methods listed to install. Is there a comparison page listing when and why should I use each method? Installing with kubeadm http://kubernetes.io/docs/getting-started-guides/kubeadm/ installing with kops http://kubernetes.io/docs/gett

Re: [kubernetes-users] How to caclulate the resource usage in a particular namespace?

Thanks, so I need to call API periodically (for example, call API every five minutes)? but what if the containers are started/shutdown during the period(let's say, five minutes? in such condition, the usage won't be calculated. I am wondering if there is any container usage events which could

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

An etcdctl cluster-health confirms on all nodes that the etcd cluster is healthy and all members are visible. The Masters are deployed using fleet so if there was a problem with etcd one of the Masters would not be running -- You received this message because you are subscribed to the Google Gr

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

It seems exactly that, it might be that the etcd are not correctly in the same cluster. I'd review that first On Sunday, November 27, 2016, 'mrpanigale' via Kubernetes user discussion and Q&A wrote: > Absolutely agree with you that this should not be happening. Unfortunately > the documentation

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

Absolutely agree with you that this should not be happening. Unfortunately the documentation does not match up with the reality once utilizing the leader elect flags on both the scheduler and the controller. Is there an etcd key within the registry hierarchy that displays the leader ip? -- You

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

The symptom you described suggests that multiple controller managers consider themselves to be leaders so they all try to create the pod (this does not have anything to do with the scheduler, which only assigns pods to nodes, it does not create pods). Specifically, if you followed the instructions

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

Yes I followed this guide as well as the coreos guide, specifically the leader elect flag which presumably solves this problem. -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To unsubscribe from this group and stop receiv

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

We have instructions for running HA masters here: http://kubernetes.io/docs/admin/high-availability/ Did you follow those? On Sun, Nov 27, 2016 at 10:25 AM, 'mrpanigale' via Kubernetes user discussion and Q&A wrote: > Each controller talking to local host API server. Each API server is an > et

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

Each controller talking to local host API server. Each API server is an etcd member of three -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To unsubscribe from this group and stop receiving emails from it, send an email

Re: [kubernetes-users] Highly available API/schedulers - strange scheduling behavior

How did you configure your multi master sharing the etcd cluster? On Sunday, November 27, 2016, 'mrpanigale' via Kubernetes user discussion and Q&A wrote: > When deploying a three node cluster with each node running all k8 > components (API, schedule, kubelet, proxy) scheduling is buggy > > When

[kubernetes-users] Highly available API/schedulers - strange scheduling behavior

When deploying a three node cluster with each node running all k8 components (API, schedule, kubelet, proxy) scheduling is buggy When starting a deployment or replication controller of 1 replica, instantly 3 pods are started before two are instantly terminated. This can lead to unexpected appli

[kubernetes-users] override default values

Hi, I want to override below values based on my VPC 10-2-64-0-19. Can some one suggest what could be the optimal value can be used? NON_MASQUERADE_CIDR="${NON_MASQUERADE_CIDR:-10.0.0.0/8}" # Traffic to IPs outside this range will use IP masquerade SERVICE_CLUSTER_IP_RANGE="${SERVICE_CLUSTER_IP_

Re: [kubernetes-users] Suited kubernetes version to production

The latest stable release is 1.4.6[0]. Please remember that there are always features and options in the production release that are marked Beta or Alpha depending on their quality. Please refer to the documentation[1]. [0]: https://github.com/kubernetes/kubernetes/releases [1]: http://kubernete

[kubernetes-users] Suited kubernetes version to production

Hi folks, I have a quick question for you :) Please, which is the more stable version of kubernetes used today that you recommend to use in production ? Cheers, Ahmed -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To u