I want to intercept the helm YAML and customize it using a Python script, and then install it. I have been doing something like helm template | python3 script... | kubectl apply -f - but of course this doesn't create a helm release in my cluster, so I lose out on helm rollback etc.
I have considered using Kustomize but it doesn't have the features that I'd like.
Is there a way to take pre-generated YAML, like that from helm template or helm install --dry-run and then install/upgrade that using helm?
Isn't that what post-renderers are for?
See https://helm.sh/docs/topics/advanced/#post-rendering
A post-renderer can be any executable that accepts rendered Kubernetes manifests on STDIN and returns valid Kubernetes manifests on STDOUT. It should return an non-0 exit code in the event of a failure. This is the only "API" between the two components. It allows for great flexibility in what you can do with your post-render process.
A post renderer can be used with install, upgrade, and template. To use a post-renderer, use the --post-renderer flag with a path to the renderer executable you wish to use:
$ helm install mychart stable/wordpress --post-renderer ./path/to/executable
I haven't used it myself yet, but it looks interesting if you want to run your own alternative kustomize.
See https://github.com/vmware-tanzu/carvel-ytt/tree/develop/examples/helm-ytt-post-renderer for an example that is not kustomize.
We have a set of kubernetes yamls which is management by kustomize and they will be deployed to different clusters. Each cluster is slightly different which means every environment will have a sub directory (environ/<envname>) including some special kustomization overwrite.
We will manually deploy new version to different environments by command kubeclt apply -k environ/env. But sometimes we do stupid thing like this: kubectl apply -k environ/env1 to the cluster env2 . So is there some method to stop doing a kubectl apply to a wrong environment?
This is a community wiki answer. Feel free to expand it.
If you are aware that you made a mistake and want to cancel the command right away than there are some options for you:
$ kill -9 $! will kill the most recent process executed by the command ($! represents its process ID).
Suspend the current process by pressing Ctrl+z and then kill it using kill -9 %% or kill -9 %+. More details regarding this approach can be found here.
EDIT:
Including the solution proposed by VASャ from the comments:
I'd use shell scripts and different configs for each cluster, like
that: deploy-cluster1.sh where I'd have kubectl --kubeconfig .kube/cluster1 apply -k environ/cluster1 or even shorter: deploy.sh env1 where deploy.sh contains: kubectl --kubeconfig .kube/$1 apply -k environ/$1
More details regarding that approach can be found here.
Recently I got a new solution here.
direnv which can change env after switch into different directories this force me to switch the KUBECONFIG env. Follow Mastering the KUBECONFIG file to get more details.
If I am using kubectx and switch kube config contexts into another cluster e.g. "Production" and run a helm uninstall, how does Helm know which cluster I am referring to?
If I run the helm list command is it only referring to what's installed on my local machine and not per Kubernetes cluster?
Helm will default to using whatever your current Kubernetes context is, as specified in the $HOME/.kube/config file.
There is standard support in the Kubernetes API libraries to read data out of this file (or an alternative specified by a $KUBECONFIG environment variable). If you're writing Go, see the documentation for the k8s.io/client-go/tools/clientcmd package. While kubectx does a bunch of things, its core uses that API to do essentially the same thing as running kubectl config use-context ....
If you want Helm to use a non-default context, there is a global option to specify it:
kubectx production
helm list
kubectx development
helm --kube-context production list
I need to pass dynamic env variable to kubectl create. Something like this
kubectl create -f app.yaml --Target=prod
Based on Target code deploys on different servers.
If you want to avoid installing 3rd party plugin then you can replace the text using sed "s/orginal/change/". It worked. I used this in Jenkins shell.
cat app.yaml | sed "s/l3-apps/l2-apps/" | kubectl create -f -
You can achieve this in two ways:
Use Helm. It is a "package manager" for Kubernetes and is built exactly for your use case (dynamic variables to configure behaviour of your resources). If it is only a single variable, "converting" your deployment is as simple as creating a new Helm chart, copy your files into templates/, modify values.yaml and use {{ .Values.target }} in your templates. See the quickstart guide for a more in-depth introduction to Helm.
If you consider Helm to be over the top for a single variable, use kubectl's capability to read from standard input. You'll need an additional templating tool (for example mustache). Rewrite your deployment to fit your templating tool. Create a dynamic data.yml in your deployment process (e.g. a simple bash script that reads from environment variables) and run something like mustache data.yml deployment.mustache | kubectl apply -f -.
kubectl config set-context allows you to configure cluster, namespace, user credentials and more and save it as a "context" in your ~/.kube/config.
The you can use --context option of kubectl exactly in a way that you used --Target in your example.
I've published a command-line tool ysed that also performs what you need.
I have Kubernetes working well in two different environments, namely in my local environment (MacBook running minikube) and as well as on Google's Container Engine (GCE, Kubernetes on Google Cloud). I use the MacBook/local environment to develop and test my YAML files and then, upon completion, try them on GCE.
Currently I need to work with each environment individually: I need to edit the YAML files in my local environment and, when ready, (git) clone them to a GCE environment and then use/deploy them. This is a somewhat cumbersome process.
Ideally, I would like to use kubectl from my Macbook to easily switch between the local minikube or GCE Kubernetes environments and to easily determine where the YAML files are used. Is there a simple way to switch contexts to do this?
You can switch from local (minikube) to gcloud and back with:
kubectl config use-context CONTEXT_NAME
to list all contexts:
kubectl config get-contexts
You can create different enviroments for local and gcloud and put it in separate yaml files.
List contexts
kubectl config get-contexts
Switch contexts
kubectl config set current-context MY-CONTEXT
A faster shortcut to the standard kubectl commands is to use kubectx:
List contexts: kubectx
Equivalent to kubectl config get-contexts
Switch context (to foo): kubectx foo
Equivalent to kubectl config use-context foo
To install on macOS: brew install kubectx
The kubectx package also includes a similar tool for switching namespaces called kubens.
These two are super convenient if you work in multiple contexts and namespaces regularly.
More info: https://ahmet.im/blog/kubectx/
If you're looking for a GUI-based solution for Mac and have the Docker Desktop installed, you can use the Docker Menu Bar icon. Here you can find "Kubernetes" menu with all the contexts you have in your kubeconfig and easily switch between them.
To get all context
C:\Users\arun>kubectl config get-contexts
To get current context
C:\Users\arun>kubectl config current-context
To switch context
C:\Users\arun>kubectl config use-context <any context name from above list>
Latest 2020 answer is here,
A simple way to switch between kubectl context,
kubectl top nodes **--context=**context01name
kubectl top nodes --context=context02name
You can also store the context name as env like
context01name=gke_${GOOGLE_CLOUD_PROJECT}_us-central1-a_standard-cluster-1
I got bored of typing this over and over so I wrote a simple bash utility to switch contexts
You can find it here https://github.com/josefkorbel/kube-switch
The canonical answer of switching/reading/manipulating different kubernetes environments (aka kubernetes contexts) is, as Mark mentioned, to use kubectl config, see below:
$ kubectl config
Modify kubeconfig files using subcommands like "kubectl config set current-context my-context"
Available Commands:
current-context Displays the current-context
delete-cluster Delete the specified cluster from the kubeconfig
delete-context Delete the specified context from the kubeconfig
get-clusters Display clusters defined in the kubeconfig
get-contexts Describe one or many contexts
rename-context Renames a context from the kubeconfig file.
set Sets an individual value in a kubeconfig file
set-cluster Sets a cluster entry in kubeconfig
set-context Sets a context entry in kubeconfig
set-credentials Sets a user entry in kubeconfig
unset Unsets an individual value in a kubeconfig file
use-context Sets the current-context in a kubeconfig file
view Display merged kubeconfig settings or a specified kubeconfig file
Usage:
kubectl config SUBCOMMAND [options]
Behind the scene, there is a ~/.kube/config YAML file that stores all the available contexts with their corresponding credentials and endpoints for each contexts.
Kubectl off the shelf doesn't make it easy to manage different kubernetes contexts as you probably already know. Rather than rolling your own script to manage all that, a better approach is to use a mature tool called kubectx, created by a Googler named "Ahmet Alp Balkan" who's on Kubernetes / Google Cloud Platform developer experiences Team that builds tooling like this. I highly recommend it.
https://github.com/ahmetb/kubectx
$ kctx --help
USAGE:
kubectx : list the contexts
kubectx <NAME> : switch to context <NAME>
kubectx - : switch to the previous context
kubectx <NEW_NAME>=<NAME> : rename context <NAME> to <NEW_NAME>
kubectx <NEW_NAME>=. : rename current-context to <NEW_NAME>
kubectx -d <NAME> [<NAME...>] : delete context <NAME> ('.' for current-context)
(this command won't delete the user/cluster entry
that is used by the context)
kubectx -h,--help : show this message
TL;DR: I created a GUI to switch Kubernetes contexts via AppleScript. I activate it via shift-cmd-x.
I too had the same issue. It was a pain switching contexts by the command line. I used FastScripts to set a key combo (shift-cmd-x) to run the following AppleScript (placed in this directory: $(HOME)/Library/Scripts/Applications/Terminal).
use AppleScript version "2.4" -- Yosemite (10.10) or later
use scripting additions
do shell script "/usr/local/bin/kubectl config current-context"
set curcontext to result
do shell script "/usr/local/bin/kubectl config get-contexts -o name"
set contexts to paragraphs of result
choose from list contexts with prompt "Select Context:" with title "K8s Context Selector" default items {curcontext}
set scriptArguments to item 1 of result
do shell script "/usr/local/bin/kubectl config use-context " & scriptArguments
display dialog "Switched to " & scriptArguments buttons {"ok"} default button 1
Cloning the YAML files across repos for different environments is definitely ideal. What you to do is templatize your YAML files - by extracting the parameters which differ from environment to environment.
You can, of course, use some templating engine and separate the values in a YAML and produce the YAML for a specific environment. But this is easily doable if you adopt the Helm Charts. To take a look at some sample charts go to stable directory at this Github repo
To take an example of the Wordpress chart, you could have two different commands for two environments:
For Dev:
helm install --name dev-release --set \
wordpressUsername=dev_admin, \
wordpressPassword=dev_password, \
mariadb.mariadbRootPassword=dev_secretpassword \
stable/wordpress
It is not necessary to pass these values on CLI though, you can store the values in a file called aptly values.yml and you could have different files for different environments
You will need some work in converting to Helm chart standards, but the effort will be worth it.
Check also the latest (docker 19.03) docker context command.
Ajeet Singh Raina ) illustrates it in "Docker 19.03.0 Pre-Release: Fast Context Switching, Rootless Docker, Sysctl support for Swarm Services"
A context is essentially the configuration that you use to access a particular cluster.
Say, for example, in my particular case, I have 4 different clusters – mix of Swarm and Kubernetes running locally and remotely.
Assume that I have a default cluster running on my Desktop machine , 2 node Swarm Cluster running on Google Cloud Platform, 5-Node Cluster running on Play with Docker playground and a single-node Kubernetes cluster running on Minikube and that I need to access pretty regularly.
Using docker context CLI I can easily switch from one cluster(which could be my development cluster) to test to production cluster in seconds.
$ sudo docker context --help
Usage: docker context COMMAND
Manage contexts
Commands:
create Create a context
export Export a context to a tar or kubeconfig file
import Import a context from a tar file
inspect Display detailed information on one or more contexts
ls List contexts
rm Remove one or more contexts
update Update a context
use Set the current docker context
Run 'docker context COMMAND --help' for more information on a command.
For example:
[:)Captain'sBay=>sudo docker context ls
NAME DESCRIPTION DOCKER ENDPOINT KUBERNETES ENDPOINT ORCHESTRATOR
default * Current DOCKER_HOST based configuration unix:///var/run/docker.sock https://127.0.0.1:16443 (default) swarm
swarm-context1
I use kubeswitch (disclaimer: I wrote the tool) that can be used just like kubectx, but is designed for a large number of kubeconfig files.
If you have to deal with hundreds or thousands of kubeconfig files, this tool might be useful to you, otherwise kubectx or kubectl config use-context might be sufficient.
For instance, it adds capabilities like reading from vault, hot reload while searching, and an index to speed up subsequent searches.
You can install it from here.
EDIT: now also includes support for GKE directly. So you can use and discover kubeconfig files without having to manually download and update them.
In case you might be looking for a simple way to switch between different contexts maybe this will be of help.
I got inspired by kubectx and kswitch scripts already mentioned, which I can recommend for most use-cases. They are helping with solving the switching task, but are breaking for me on some bigger or less standard configurations of ~/.kube/config. So I created a sys-exec invocation wrapper and a short-hand around kubectl.
If you call k without params you would see an intercepted prompt to switch context.
Switch kubectl to a different context/cluster/namespace.
Found following options to select from:
>>> context: [1] franz
>>> context: [2] gke_foo_us-central1-a_live-v1
>>> context: [3] minikube
--> new num [?/q]:
Further, k continues to act as a short-hand. The following is equivalent:
kubectl get pods --all-namespaces
k get pods -A
k p -A
yes, i think this is what your asking about. To view your current config, use kubectl config view. kubectl loads and merges config from the following locations (in order)
--kubeconfig=/path/to/.kube/config command line flag
KUBECONFIG=/path/to/.kube/config env variable
$HOME/.kube/config - The DEFAULT
i use --kubeconfig since i switch alot between multiple clusters. its slightly cumbersome but it works well.
see these for more info.
https://kubernetes.io/docs/tasks/administer-cluster/share-configuration/ and https://kubernetes.io/docs/concepts/cluster-administration/authenticate-across-clusters-kubeconfig/