[kubernetes-users] Re: Is it possible to set up a K8s cluster federation with two local clusters?

Thanks Madhusudan for the sharing which is very helpful!!! And I have a few more questions: 1. I tried to use one of my on-prem clusters as host cluster to deploy federation control plane via "kubefed init", and I found federation-controller-manager failed to start with the error below: F0302

Re: [kubernetes-users] Can I add side car container to a running pod?

Afaik, is not possible. And the idea is that deploys are easy, so you do them a lot, and not mutate your things but just deploy (to add a sidecar container or whatever). But maybe there is a way that I don't know about On Wednesday, March 1, 2017, Tamal Saha wrote: > Hi, >

[kubernetes-users] Can I add side car container to a running pod?

Hi, I am working on a TPR controller that will add a side card container to a pod. From what I can tell so far is that a side car container can't be added to an existing pod. I get an error message like below: 1. pods "two-containers" was not valid: 2. # *

[kubernetes-users] Re: Why isn't my DaemonSet rescheduling my Evicted pod?

Oh, that makes sense. Thanks for fixing this; I'm looking forward to 1.6! On Wednesday, March 1, 2017 at 10:37:10 AM UTC-8, Janet Kuo wrote: > > The short answer is that it's going to be fixed in Kubernetes 1.6. The > workaround is to manually kill the failed pod, so that DaemonSet controller >

Re: [kubernetes-users] security for Python micro-PaaS on GKE

You can consider having different clusters for different things, too. But network isolation seems critical and I don't think all ways to do networking today support the kubernetes network policy. I'd have that in mind when choosing the way to do networking (besides all that post says). On

[kubernetes-users] Re: Why isn't my DaemonSet rescheduling my Evicted pod?

The short answer is that it's going to be fixed in Kubernetes 1.6. The workaround is to manually kill the failed pod, so that DaemonSet controller can recreate it. Because DaemonSet controller doesn't schedule pods through scheduler

Re: [kubernetes-users] Kubernetes Volume mountpath

you have to look at your logs or something. If you're saying the files are identical in both cases, then I don't have a starting place. On Wed, Mar 1, 2017 at 9:52 AM, Montassar Dridi wrote: > it's the same war file, my application worked fine without volume but as >

Re: [kubernetes-users] Kubernetes Volume mountpath

it's the same war file, my application worked fine without volume but as soon as I mountpath: /usr/local/tomcat/webapps I get that error even though all the subdirectories are there not empty and have all their folders and files but but when i mountpath to a subdirectory in

Re: [kubernetes-users] Kubernetes Volume mountpath

I find that `kubectl exec`ing in and poking around often illuminates why it didn't do what I expected. E.g. permissions (if docker makes a dir they can be wrong, for example). On Wed, Mar 1, 2017 at 9:45 AM, Rodrigo Campos wrote: > If When using no volumes your app returns

Re: [kubernetes-users] Kubernetes Volume mountpath

If When using no volumes your app returns 404, then is something in your app. You can try to run the container locally. But really nothing else we can do, if the app doesn't work as expected just when running the container, you need to understand the app layer and see why it doesn't do what you

Re: [kubernetes-users] Kubernetes Volume mountpath

Truthfully, I don't know enough about tomcat - with nginx or apache, I would expect a 404 if you don't have an index.html, for example. On Wed, Mar 1, 2017 at 9:20 AM, Montassar Dridi wrote: > that make sense > can you explain this to me and sorry for taking this long

[kubernetes-users] security for Python micro-PaaS on GKE

I'm building a Python-only micro-PaaS (at least that's what I'm calling it ;)) running on GKE/Kubernetes - something similar to PythonAnywhere, but for a specific business niche. I'm looking for some advice around the security of the system. In addition to following the Kubernetes security