kubesphere-sigs / ks Goto Github PK

View Code? Open in Web Editor NEW

22.0 3.0 16.0 477 KB

ks is a tool that makes it be easy to work with KubeSphere.

License: MIT License

Makefile 0.68% Go 96.96% Dockerfile 0.07% Groovy 2.29%

kubesphere kubectl kubectl-command kubectl-plugins krew krew-plugin hacktoberfest

ks's Introduction

ks

ks is a tool which makes it be easy to work with KubeSphere.

Get started

Install it via hd:

curl -L https://github.com/linuxsuren/http-downloader/releases/latest/download/hd-linux-amd64.tar.gz | tar xzv
sudo mv hd /usr/bin/hd
hd install kubesphere-sigs/ks

Features

All features below work with KubeSphere instead of other concept.

Component Management
- Enable (disable) components
- Update a component manually or automatically
- Output the logs of a KubeSphere component
- Edit a KubeSphere component
- Describe a KubeSphere component (wrapper of kubectl describe)
Pipeline management
- Create a Pipeline with java, go template
- Edit a Pipeline without give the fullname (namespace/name)
User Management
Working with k3s via setting environment kubernetes_type=k3s
KubeSphere installation
- Install KubeSphere via ks-installer
- Install KubeSphere via k3d
- Install KubeSphere via kubekey
- Install KubeSphere via kind

Component

➜  ~ kubectl ks com
Manage the components of KubeSphere

Usage:
  ks component [command]

Aliases:
  component, com

Available Commands:
  edit        Edit the target component
  enable      Enable or disable the specific KubeSphere component
  exec        Execute a command in a container.
  kill        Kill the pods of the components
  log         Output the log of KubeSphere component
  reset       Reset the component by name
  watch       Update images of ks-apiserver, ks-controller-manager, ks-console

Pipeline

➜  ~ kubectl ks pip
Usage:
  ks pipeline [flags]
  ks pipeline [command]

Aliases:
  pipeline, pip

Available Commands:
  create      Create a Pipeline in the KubeSphere cluster
  delete      Delete a specific Pipeline of KubeSphere DevOps
  edit        Edit the target pipeline
  view        Output the YAML format of a Pipeline

Flags:
  -h, --help   help for pipeline

Use "ks pipeline [command] --help" for more information about a command.

Installation

Install KubeSphere with kind or kk

Usage:
  ks install [command]

Available Commands:
  kind        Install KubeSphere with kind
  kk          Install KubeSphere with kubekey (aka kk)

ks's People

Contributors

Stargazers

Watchers

Forkers

yjuns foreverif johnniang linuxsuren-bot linuxsuren shihaoh mangogoforward yk199316 wst021 kubesphere-community selamanse jzhupup yudong2015 loongson-cloud-community tianyuanxixue caiqi1111

ks's Issues

Request to support component uninstallation feature

At present, official website has provided a document about how to uninstall each component.

But it's not very friendly for end users to execute so many command at the same time.

Therefore, I request to support component uninstallation in ks command. e.g.:

ks uninstall devops
ks uninstall logging
ks uninstall alerting
...

# Or uninstall all component but basic component
ks uninstall all

/kind feature

In k3d v5.0.0+ , not compatible

k3d version: v5.0.0
exec ks install k3d
Action:

FATA[0000] failed to transform ports: Failed to parse node filters: invalid format or empty subset in 'agent[0]'

This is how the k3d documentation describes it

 -p, --port [HOST:][HOSTPORT:]CONTAINERPORT[/PROTOCOL][@NODEFILTER]   Map ports from the node containers (via the serverlb) to the host (Format: [HOST:][HOSTPORT:]CONTAINERPORT[/PROTOCOL][@NODEFILTER])
                                                                        - Example: `k3d cluster create --agents 2 -p 8080:80@agent:0 -p 8081@agent:1`

Add support to create/trigger Pipeline

It's very useful in some cases:

Run some quick tests against Pipeline

Add support to install KubeSphere in kind

This feature might be a plugin of this repo.

cat <<EOF | kind create cluster --image kindest/node:v1.18.2 --config=-
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
  kubeadmConfigPatches:
  - |
    kind: InitConfiguration
    nodeRegistration:
      kubeletExtraArgs:
        node-labels: "ingress-ready=true"
  extraPortMappings:
  - containerPort: 30880
    hostPort: 30881
    protocol: TCP
  - containerPort: 30180
    hostPort: 30181
    protocol: TCP
EOF

docker pull kubesphere/ks-installer:v3.0.0
docker pull kubesphere/ks-apiserver:v3.0.0
docker pull kubesphere/ks-controller-manager:v3.0.0
docker pull kubesphere/ks-console:v3.0.0
docker pull redis:5.0.5-alpine
docker pull osixia/openldap:1.3.0
docker pull minio/minio:RELEASE.2019-08-07T01-59-21Z
docker pull mysql:8.0.11
kind load docker-image kubesphere/ks-installer:v3.0.0
kind load docker-image kubesphere/ks-apiserver:v3.0.0
kind load docker-image kubesphere/ks-controller-manager:v3.0.0
kind load docker-image kubesphere/ks-console:v3.0.0
kind load docker-image redis:5.0.5-alpine
kind load docker-image osixia/openldap:1.3.0
kind load docker-image minio/minio:RELEASE.2019-08-07T01-59-21Z
kind load docker-image mysql:8.0.11

kubectl cluster-info --context kind-kind
kubectl apply -f https://github.com/kubesphere/ks-installer/releases/download/v3.0.0/kubesphere-installer.yaml
kubectl apply -f https://github.com/kubesphere/ks-installer/releases/download/v3.0.0/cluster-configuration.yaml

# load the necessary images for devops
docker pull kubesphere/jenkins-uc:v3.0.0
docker pull jenkins/jenkins:2.176.2
kind load docker-image kubesphere/jenkins-uc:v3.0.0
kind load docker-image jenkins/jenkins:2.176.2

# load the image of Jenkins agent builder
docker pull jenkins/jnlp-slave:3.27-1
docker pull kubesphere/builder-base:v2.1.0
docker pull kubesphere/builder-nodejs:v2.1.0
docker pull kubesphere/builder-go:v2.1.0
docker pull kubesphere/builder-maven:v2.1.0
kind load docker-image kubesphere/builder-base:v2.1.0
kind load docker-image kubesphere/builder-nodejs:v2.1.0
kind load docker-image kubesphere/builder-go:v2.1.0
kind load docker-image kubesphere/builder-maven:v2.1.0
kind load docker-image jenkins/jnlp-slave:3.27-1

kubectl ks com enable devops

Record a video to show the usage of this project

asciinema is a wonderful tool that can help with that. Two big pros:

Easy to use: asciinema rec
Allow users to upload the video files to https://asciinema.org/. That means no waste storage of the git repository

Example

https://github.com/rancher/k3d/blob/main/README.md#k3s-in-docker

/good-first-issue

No root permission, no permission to install k3d

When I use not root permission and install k3d, tips no permission. but when the user has ![image](https://user-images.githubusercontent.com/20575981/132883990-564cbb88-73e2-4ec0-81ff-a0991b8cfbc8.png)do permission, sudo ks install k3d is ok.

So I suggest describing the installation method clearly.

Request to add event list for Pipeline, PipelineRun in Pipeline dashboard

Currently, If we encounter a problem with PipelineRun, we have no idea to know what happened. But our reconcilers may store some events about the result of Pipeline or PipelineRun.

As a result, I request to add event list for Pipeline, PipelineRun in Pipeline dashboard to let users know what happened just now whether it is successful.

/kind feature

Add a sub-command to install kubekey

If there's no KubeSphere in the current cluster, it would be good to install it via kk.

Couldn't install kubectl-ks correctly via README.md

In README.md, I was trying to execute command hd install -t 8 linuxsuren/ks/kubectl-ks to install k8s and ks, but failed eventually. Full error log as follow:

start to fetch the config
From https://github.com/LinuxSuRen/hd-home
 * branch            master     -> FETCH_HEAD
HEAD is now at 48ee4a8 Add cgit in the package list (#27)
start to download from https://github.com/linuxsuren/ks/releases/latest/download/kubectl-ks-windows-amd64.tar.gz
Error: cannot download from https://github.com/linuxsuren/ks/releases/latest/download/kubectl-ks-windows-amd64.tar.gz, response error: failed to download from 'https://github.com/linuxsuren/ks/releases/latest/download/kubectl-ks-windows-amd64.tar.gz': status code: 404, content length error: <nil>

I found that github releases' detail didn't list any assets which name started with "kubectl-ks".

Request to release a new version 0.0.54

Since v0.0.53 was released, several bugs have been fixed.

I'm looking forward to the next release to test KubeSphere with pre-release version.

/cc @kubesphere-sigs/sig-devops

Deploy svn for the test purpose

docker run -d --name svn-server -p 80:80 -p 3690:3690 elleflorio/svn-server

Add support to reset components to the latest nightly build

This is an example of usage:

kubectl ks com reset --version nightly

Add support install KubeSphere on containerd

References

https://kubesphere.com.cn/forum/d/3054-dockerkubernetes (Chinese)

Add support to add a k8s context into kubeconfig

Add completion support for command 'ks com watch'

Add support create docker registry repo in secret format

Below is an example of docker registry as secret in KubeSphere project:

apiVersion: v1
data:
  .dockerconfigjson: eyJhdXRocyI6eyJodHRwczovL2RvY2tlci5pbyI6eyJ1c2VybmFtZSI6InN1cmVucGkiLCJwYXNzd29yZCI6IjEyMzQ1NiIsImVtYWlsIjoiIiwiYXV0aCI6ImMzVnlaVzV3YVRveE1qTTBOVFk9In19fQ==
kind: Secret
metadata:
  annotations:
    kubesphere.io/creator: admin
  creationTimestamp: "2021-03-09T01:07:50Z"
  name: "22"
  namespace: b2i
  resourceVersion: "240359"
  selfLink: /api/v1/namespaces/b2i/secrets/22
  uid: a6f09f64-c3b2-46e0-b898-91b70bc409ad
type: kubernetes.io/dockerconfigjson

Why need this?

It can benefit automated tests.

GC Pipelinerun not work with days_to_keep and num_to_keep in Pipeline

GC Pipelineruns that outdated pipelinerun base on discarder inPipeline not be deleted.

the configuration: num_to_keep is 2, but there are 4 records(pipelinerun).

Add support install KubeSphere in k3d

k3d is a very useful tool when you try some new features of KubeSphere.

You can start a k3s cluster via: k3d cluster create -p "30880:30880@agent[0]" --agents

Add a timeout of ks component watch command

The default timeout could be one hour.

Cannot use ks in the KubeSphere kubectl console

You will get the errors like below:

/ # ks pip
stat /root/.kube/config: no such file or directory
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x18 pc=0x1387442]

goroutine 1 [running]:
github.com/linuxsuren/ks/kubectl-plugin/pipeline.getAllNamespace(0x0, 0x0, 0x203000, 0xc0000b3958, 0x759cf3cb6dc993e)
        /home/runner/work/ks/ks/kubectl-plugin/pipeline/root.go:64 +0x82
github.com/linuxsuren/ks/kubectl-plugin/pipeline.getNamespace(0x0, 0x0, 0x250e990, 0x0, 0x0, 0x0, 0x153de60, 0xffffffffffffffff, 0xc0
000b3a20)
        /home/runner/work/ks/ks/kubectl-plugin/pipeline/root.go:44 +0xd5
github.com/linuxsuren/ks/kubectl-plugin/pipeline.getPipelines(0x0, 0x0, 0x250e990, 0x0, 0x0, 0xc0000b3b90, 0xc0000b3bb8, 0xc0000b3bb8
, 0x5bad44, 0xc0000e3700, ...)
        /home/runner/work/ks/ks/kubectl-plugin/pipeline/edit.go:95 +0x3f2
github.com/linuxsuren/ks/kubectl-plugin/pipeline.NewPipelineCmd.func1(0xc0001458c0, 0x250e990, 0x0, 0x0, 0x0, 0x0)
        /home/runner/work/ks/ks/kubectl-plugin/pipeline/root.go:21 +0x6c
github.com/spf13/cobra.(*Command).execute(0xc0001458c0, 0x250e990, 0x0, 0x0, 0xc0001458c0, 0x250e990)
        /home/runner/go/pkg/mod/github.com/spf13/[email protected]/command.go:850 +0x460
github.com/spf13/cobra.(*Command).ExecuteC(0xc0000dd080, 0x7, 0x174ce88, 0x1)
        /home/runner/go/pkg/mod/github.com/spf13/[email protected]/command.go:958 +0x349
github.com/spf13/cobra.(*Command).Execute(...)
        /home/runner/go/pkg/mod/github.com/spf13/[email protected]/command.go:895
github.com/linuxsuren/go-cli-alias/pkg/cmd.Execute(0xc0000dd080, 0x175144d, 0x7, 0xc000375400, 0x13, 0x13, 0x0)
        /home/runner/go/pkg/mod/github.com/linuxsuren/[email protected]/pkg/cmd/root.go:130 +0x644
main.main()
        /home/runner/work/ks/ks/main.go:53 +0x32b

Add support install kubesphere

The usage could be like this: kubectl ks install

Add support enable multiple-login with KubeSphere

Please take the following steps:

run command ks -n kubesphere-system edit cm kubesphere-config then change authentication.multipleLogin to true
restart ks-apiserver

Failed to exec into a component

Please see also the following error output:

kubectl exec [POD] [COMMAND] is DEPRECATED and will be removed in a future version. Use kubectl exec [POD] -- [COMMAND] instead.

Request to support installation of KubeSphere v3.1.1

[root@node1 ~]# ks install kk --components devops --version v3.1.1

Flags:
      --components stringArray   The components which you want to enable after the installation
      --container string         Indicate the container runtime type. Supported: docker, containerd (default "docker")
      --fetch                    Indicate if fetch the latest config of tools (default true)
  -h, --help                     help for kk
  -v, --version string           The version of KubeSphere. Support value could be v3.1.0, nightly, nightly-20210309. nightly equals to nightly-latest (default "v3.1.0")
      --zone string              Set environment variables, for example export KKZONE=cn (default "cn")

not support version: v3.1.1

Request to support import a member cluster

Possible usage: ks opt cluster import

This command might do the following things:

check the current cluster if is the host role
let users choose from context (~/.kube/config)
import

Can't add service when deleted metadata by advanced settings

When I deleted the metadata by advanced settings, I can't add the service. The default value is like this:

but when I deleted the metadata, tips the label can not null, and checkbox of Add metadata can not cancel selection

Request to install multiple clusters which has the host and members

Possible usage: ks install k3d --role-host 1 --role-member 2

This feature might rely on #215

This command can install multiple KubeSphere clusters which based on k3d.

Incorrect GC behaviour for PipelineRuns

What's the problem?

Recently, I encountered that my pending or running PipelineRuns disappeared suddenly, even in Jenkins. So I guess it may be GCed by cronjob-gc. I will show the evidence up below.

Steps to reproduce

Create a simple Pipeline with Jenkinsfile as below:

pipeline {
    agent any 
    stages {
        stage('Greet') { 
            steps {
              sh 'echo "Hello, Jenkins!" && ping bing.com'
            }
        }
    }
}

Click Run button to run the Pipeline
Then Run the GC job immediately

kubectl -n kubesphere-devops-system create job gc-test-job --from=cronjob/devops

See the PipelineRun you just run

/kind bug

Add diagnose support for KubeSphere

Some potential issues might be able to detect by a CLI. For example, s2ibuildertemplates is missing.

Failed to install KubeSphere via ks install kk on a fresh machine

The command is:

ks install kk --kubernetesVersion v1.21.5

The error output is:

time="10:40:52 CST" level=error msg="Failed to connect to 192.168.0.3: Failed to validate ssh connection parameters: Failed to read keyfile \"/root/.ssh/id_rsa\": open /root/.ssh/id_rsa: no such file or directory" node=192.168.0.3
time="10:40:52 CST" level=warning msg="Task failed ..."
time="10:40:52 CST" level=warning msg="error: interrupted by error"
Error: Failed to init OS: interrupted by error
Failed to init OS: interrupted by error

It still cannot work after I generated the ssh key with the following command:

ssh-keygen

the error output is:

time="10:57:28 CST" level=error msg="Failed to connect to 192.168.0.3: could not establish connection to 192.168.0.3:22: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain" node=192.168.0.3
time="10:57:28 CST" level=warning msg="Task failed ..."
time="10:57:28 CST" level=warning msg="error: interrupted by error"
Error: Failed to init OS: interrupted by error
Failed to init OS: interrupted by error

then, I put the content of /root/.ssh/id_rsa.pub into the end of /root/.ssh/authorized_keys. It works well.

Suggestion

I'd like to suggest doing all those work by the command ks install kk. So users don't need to do it over and over again.

Add gray release support

You can get the release strategy list via: ks get --all-namespaces strategy. Below is an example:

apiVersion: servicemesh.kubesphere.io/v1alpha2
kind: Strategy
metadata:
  annotations:
    app.kubernetes.io/icon: /assets/bookinfo.svg
    kubesphere.io/creator: admin
    servicemesh.kubesphere.io/newWorkloadName: details-v2
    servicemesh.kubesphere.io/oldWorkloadName: details-v1
    servicemesh.kubesphere.io/workloadReplicas: "1"
    servicemesh.kubesphere.io/workloadType: deployments
  creationTimestamp: "2021-03-15T05:43:22Z"
  generation: 1
  labels:
    app: details
    app.kubernetes.io/name: bookinfo
    app.kubernetes.io/version: v1
  name: blue-test-1
  namespace: ks
  ownerReferences:
  - apiVersion: app.k8s.io/v1beta1
    blockOwnerDeletion: true
    controller: false
    kind: Application
    name: bookinfo
    uid: 8eb1bbec-d44f-4d5f-8507-500e92775234
  resourceVersion: "152221"
  selfLink: /apis/servicemesh.kubesphere.io/v1alpha2/namespaces/ks/strategies/blue-test-1
  uid: 2ca3bc16-1e91-4cc5-b200-e9630bd2fac0
spec:
  governor: v1
  hosts: details
  principal: v1
  protocol: http
  selector:
    matchLabels:
      app: details
      app.kubernetes.io/name: bookinfo
      app.kubernetes.io/version: v1
  strategyPolicy: WaitForWorkloadReady
  template:
    spec:
      hosts:
      - details
      http:
      - route:
        - destination:
            host: details
            subset: v1
          weight: 100
        - destination:
            host: details
            subset: "1"
          weight: 0
  type: Bluegreen

Deploy bitbucket for the test purpose

The possible docker images:

atlassian/bitbucket-server:7.10.0
atlassian/bitbucket-server:7.6.4
atlassian/bitbucket-server:7.2.4

Get started

Run it via:

docker volume create --name bitbucketVolume
docker run -d -v bitbucketVolume:/var/atlassian/application-data/bitbucket \
    -p 30000:7990 -p 30001:7999 \
    atlassian/bitbucket-server:7.10.0

Add a monkey plugin to manipulate KubeSphere randomly

go-cli-plugin is a library for go cli plugin.

Why pipelinerun gc ignores multi-branch-pipeline?

Add support to submit issues of KubeSphere

A possible usage could be:

kubectl ks issue create --title title --body body --label label --label label-1 --project devops

Add a command to setup KubeSphere

case 1

Sometimes we need multiple environments for development and testing purposes. It's hard to memorize the address of those KubeSphere clusters. Even worse, part of those environments might be temporary. Give that, it would be good to have a simple way to set up a remote kubeclt connection locally, open the console of the KubeSphere cluster.

You can consider the following usage:

ks config add --cluster dev --host 192.168.0.3

The above command will get the token of the default serviceAccount, then add it to ~/.kube/config as dev. Secondly, create a configMap to store the address of the cluster. It could be like:

kind: ConfigMap
apiVersion: v1
data:
  externalAddress: http://192.168.0.3:30880

If you want to open it in your browser, just need to type: ks open dev.

Add support to create a project in KubeSphere workspace

Below is an example of a project:

apiVersion: v1
kind: Namespace
metadata:
  annotations:
    kubesphere.io/creator: admin
  finalizers:
  - finalizers.kubesphere.io/namespaces
  labels:
    kubesphere.io/namespace: b2i
    kubesphere.io/workspace: test
  name: b2i
  ownerReferences:
  - apiVersion: tenant.kubesphere.io/v1alpha1
    blockOwnerDeletion: true
    controller: true
    kind: Workspace
    name: test
    uid: 45ed0813-fb91-41d7-af5c-1aa1526d99d4
spec:
  finalizers:
  - kubernetes
status:
  phase: Active

Why need this?

It can benefit automated tests.

ks app update command report error: "failed to push changes, error is: failed to push branch, error is: failed to commit changes, error is author field is required"

I use ks app update command, as follows:

ks app update --app-name test11111 --app-namespace testqgg2s --name old-image-addr --newName new-image-addr:newtag --git-password 'O>hx@q012I' --git-username=app-mwcn --git-target-branch dev

encountered an error:

failed to push changes, error is: failed to push branch, error is: failed to commit changes, error is author field is required

I checked the source code and probably found the code location where the problem is：

Add a link of Unix philosophy to guide us for the command line program

It's very important to have rules for the programming and code review.

References

I learned those links from @cndoit18

Add support to maintain the components of ks-devops

Add support to set federal cluster

KubeSphere support the federal cluster. It can set a Kubernetes cluster to be a member or host role. So, the first feature that I want to have is making a cluster to be a member or host.

Second, support to add a cluster to be a member of the host cluster. It can be a direct connection or agent mode.

Add support to pre-load images in KubeSphere cluster

Some images will be loaded during the runtime instead of the installation phase of KubeSphere. It's very reasonable. But for the users, might need to wait for a while before all the necessary images are ready. Case 2, users might want to use KubeSphere in a private environment which means there's no internet.

So what I want to propose is that provide a command to pre-load all necessary images as the need to the target environment.

For the machine which can access the internet, the usage could be ks images load --components devops

For no internet case, the usage could be ks images save --components devops, then load them by offline mode.

How to get the whole images list

You can get the list from file images-list.txt

One question is that it does not tell us the relation between images and components. I guess re-org the file from .txt to .yaml will be better.

Username & Password

apiVersion: v1
data:
  password: YXNkZg==
  username: YXNkZg==
kind: Secret
metadata:
  annotations:
    credential.devops.kubesphere.io/autosync: "true"
    credential.devops.kubesphere.io/datahash: 848fbcbc7d
    credential.devops.kubesphere.io/syncstatus: successful
    credential.devops.kubesphere.io/synctime: 2021-03-08 16:45:31.889544672 +0800
      CST m=+6505.581589688
    kubesphere.io/creator: admin
  creationTimestamp: "2021-03-08T08:45:31Z"
  finalizers:
  - finalizers.kubesphere.io/credential
  labels:
    app: abc
  name: abc
  namespace: test2tkk7
  resourceVersion: "31863"
  selfLink: /api/v1/namespaces/test2tkk7/secrets/abc
  uid: a05fd2b8-2661-4df5-811d-ac4252817205
type: credential.devops.kubesphere.io/basic-auth

Add support to use ks image in the kubectl admin of KubeSphere

Currently, you can use the following command to archive that. But I believe that it can be easier.

ks -n kubesphere-controls-system set image deploy kubectl-admin kubectl=surenpi/ks:v0.0.26

For example, we can use do it like: ks config use-ks

kubesphere-sigs / ks Goto Github PK

ks's Introduction

ks

Get started

Features

Component

Pipeline

Installation

ks's People

Contributors

Stargazers

Watchers

Forkers

ks's Issues

Example

References

Why need this?

What's the problem?

Steps to reproduce

Suggestion

Get started

case 1

Why need this?

References

How to get the whole images list

Username & Password

Recommend Projects

Recommend Topics

Recommend Org