harvester-auto

Create a Harvester Cluster with a single slack command.

Features

• Multi-User
• Multi-Namespace
• Multi-Cluster
• Multi-Node-Per-Cluster
• Multi-Disk-Per-Node
• Multi-NIC-Per-Node
• Multi-Repo
• Multi-Branch
• Auto-Download-Release
• Auto-Build-ISO
• Auto-Deploy-Cluster
• Auto-Init-Cluster
• Auto-Build-Push-UI
• Auto-Build-Push-Patch-Image
• Auto-Notification
• Auto-Upgrade
• Auto-Test

Architecture

Environment

Node Setup

Vagrant Libvirt

sudo apt purge vagrant-libvirt
sudo apt-mark hold vagrant-libvirt
sudo apt update
sudo apt install -y qemu libvirt-daemon-system ebtables libguestfs-tools vagrant ruby-fog-libvirt
vagrant plugin install vagrant-libvirt

Create 4 pools to store virtual disks. In order to improve performance and stability, it is recommended that different pools correspond to different underlying physical nvme disks

create_pool(){
	local pool_name=$1
	local pool_target=$2
	virsh pool-define-as $1 dir --target $2
	virsh pool-autostart $1
	virsh pool-start $1
}
create_pool pool1 /var/lib/libvirt/images/pool1
create_pool pool2 /var/lib/libvirt/images/pool2
create_pool pool3 /var/lib/libvirt/images/pool3
create_pool pool4 /var/lib/libvirt/images/pool4

Docker

curl -sL https://releases.rancher.com/install-docker/20.10.sh | bash -
sudo systemctl enable --now docker

Proxy

Since the Harvester nodes created use a private network, all are only accessible on the host node. In order to access the Harvester UI remotely and use kubectl to manage the cluster, running a socks5 proxy server on the host

# refer to https://github.com/serjs/socks5-server
sudo docker run -d --name socks5 --restart=unless-stopped -p 1080:1080 serjs/go-socks5-proxy

NoVNC

sudo snap install novnc

WebSSH

# refer to https://github.com/huashengdun/webssh
sudo docker run -d --name wssh --restart=unless-stopped -p 8888:8888 futuretea/wssh

WebTail

# refer to https://github.com/LeKovr/webtail
logs_dir="/workspace/logs"
sudo docker run -d --name webtail --restart=unless-stopped -p 8080:8080 -v ${logs_dir}:/mnt ghcr.io/lekovr/webtail --root /mnt

Nodejs

Use Nodejs to build the UI

• Refer to the documentation https://computingforgeeks.com/install-node-js-14-on-ubuntu-debian-linux/ to install nodejs

Tools

Some tools are required to run the scripts.

sudo apt install -y ansible sshpass jq
sudo pip install jinja2-cli
sudo snap install yq
sudo snap install task --classic
# kubectl
curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
chmod +x kubectl
mv kubectl /usr/local/bin/
# helm
curl -fsSL -o get_helm.sh https://raw.githubusercontent.com/helm/helm/main/scripts/get-helm-3
chmod 700 get_helm.sh
./get_helm.sh
# terraform
wget -O- https://apt.releases.hashicorp.com/gpg | gpg --dearmor | sudo tee /usr/share/keyrings/hashicorp-archive-keyring.gpg
echo "deb [signed-by=/usr/share/keyrings/hashicorp-archive-keyring.gpg] https://apt.releases.hashicorp.com $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/hashicorp.list
sudo apt update && sudo apt install terraform

Dependent service

Use the following LAN services to speed up the build process.

MinIO

Use MinIO to serve the built ISO files.

• Refer to the documentation https://min.io/download to install minio
• create a harvester-iso bucket
• download minio client mc

wget https://dl.min.io/client/mc/release/linux-amd64/mc
chmod +x mc
mv mc /usr/local/bin/

• set alias

mc alias set myminio <minio url> <minio access key> <minio secret key>

• set policy

mc anonymous set download myminio/harvester-iso

Harbor

Use Harbor to serve the built docker images.

• Refer to the documentation https://goharbor.io/ to install Harbor
• create a rancher project
• docker login

docker login <Harbor domain>

Usage

Preparing your Slack App

Refer to https://github.com/shomali11/slacker#preparing-your-slack-app

Clone

git clone https://github.com/futuretea/harvester-auto.git
cd harvester-auto

Configure

Change the dns nameserver address and image repo configurations in ./commands/_config.sh

cd commands
cp _config.sh.example _config.sh
vim _config.sh
cd -

Fill in the Slack app token and user configurations in ./configs/config.yaml

cd configs
cp config.yaml.example config.yaml
vim config.yaml
cd -

Build

task go:build
mv ./bin/harvester-auto .

Run for testing

./harvester-auto

Run in background

cat > /tmp/harvester-auto.service <<EOF
[Unit]
Description=Harvester Auto Service
After=network.target

[Service]
Type=simple
User=${USER}
Restart=on-failure
RestartSec=5s
ExecStart=${PWD}/harvester-auto
WorkingDirectory=${PWD}

[Install]
WantedBy=multi-user.target
EOF
sudo cp /tmp/harvester-auto.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable --now harvester-auto

Update

task update

Send commands from slack

Send help to the Slack app to get the help message

PRs format in commands

1. A number 0 means no PRs
2. A number greater than 0 means the PR number, e.g. 123 means the PR 123
3. A string means a branch name in fork repo, e.g. futuretea:dev means the branch dev in user futuretea's fork repo
4. Support multiple PRs or branches, separated by comma, e.g. 123,456,futuretea:dev
5. Support backport PRs, e.g. [email protected] means backport PR 123 to branch v1.1

A complete example: 123,futuretea:[email protected] will:

• Checkout the base branch, default to master
• Create a base branch named pr-123-futuretea-dev-v1-1 from the base master
• Backport PR 123 to branch v1.1, cherry-pick the changes
• Backport branch dev in user futuretea's fork repo to branch v1.1, cherry-pick the changes

Commands

• help - help

• ping - Ping! *

  Example: ping

• history historyNumber - Show history *

  Example: history

  Example: history 100

• l - List Harvester clusters *

  Example: l

• c clusterID - Show/Set Current Harvester cluster *

  Example: c (show current cluster id)

  Example: c 1 (set current cluster id to 1)

• status - Show Harvester cluster status *

  Example: status

• url - Show Harvester cluster URLs *

  Example: url

• version - Show Harvester version *

  Example: version

• name name - Show/Set Harvester name *

  Example: name

  Example: name test-cluster

• settings - Show Harvester settings *

  Example: settings

• pis kubeNamespace - Show Pod Images *

  Example: pis

  Example: pis cattle-system

• pods kubeNamespace - Show Pods *

  Example: pods

  Example: pods cattle-system

• vms kubeNamespace - Show VMs *

  Example: vms

  Example: vms harvester-public

• settings - Show Harvester settings *

  Example: settings

• kubeconfig - Show Harvester cluster kubeconfig content *

  Example: kubeconfig

• sshconfig - Show ssh config for connecting *

  Example: sshconfig

• destroy - Destroy Harvester cluster nodes *

  Example: destroy

• start - Start Harvester cluster nodes *

  Example: start

• restart - Restart Harvester cluster nodes *

  Example: restart

• stop - Stop Harvester cluster nodes *

  Example: stop

• snaps - List Harvester cluster snapshots *

  Example: snaps

• snap - Snapshot Harvester cluster nodes *

  Example: snap 1node

  Example: snap 3node

• revert - Revert Harvester cluster nodes *

  Example: revert 1node

  Example: revert 3node

• virsh command args - virsh command warpper *

  Example: virsh list

• ps - Show running jobs *

  Example: ps

• log job lineNumber - Tail Job logs *

  Example: log 2c

  Example: log 2c 100

  Example: log 2ui

  Example: log 2ui 100

  Example: log 2pt

  Example: log 2pt 100

  Example: log 2iso

  Example: log 2iso 100

  Example: log sc

  Example: log sc 100

  Example: log up

  Example: log up 100

• kill job - Kill running job *

  Example: kill 2c

  Example: kill 2pt

  Example: kill 2iso

  Example: kill 2ui

  Example: kill sc

  Example: kill up

• pr2iso harvesterPRs harvesterInstallerPRs - Build Harvester iso after merging PRs or checkout branches *

  Example: pr2iso 0 0

• pr2c harvesterPRs harvesterInstallerPRs harvesterConfigURL - Create a Harvester cluster after merging PRs or checkout branches, always build ISO *

  Example: pr2c 0 0

• pr2cNoBuild harvesterPRs harvesterInstallerPRs harvesterConfigURL - Create a Harvester cluster based on PRs or branches, but use the built ISO from pr2c *

  Example: pr2cNoBuild 0 0

• v2c harvesterVersion harvesterConfigURL - Create a Harvester cluster after downloading the ISO *

  Example: v2c v1.1.1

• pr2up harvesterPRs harvesterInstallerPRs - Upgrade a Harvester cluster after merging PRs or checkout branches, always build ISO *

  Example: pr2up 0 0

• pr2upNoBuild harvesterPRs harvesterInstallerPRs - Upgrade a Harvester cluster based on PRs or branches, but use the built ISO from pr2c *

  Example: pr2upNoBuild 0 0

• v2up harvesterVersion - Upgrade a Harvester cluster after downloading the ISO *

  Example: v2up v1.1.2-rc5

• scale nodeNumber - Scale Harvester nodes *

  Example: scale 2

  Example: scale 3

• pr2ui uiPRs - Build Harvester Dashboard *

  Example: pr2ui 0

• pr2pt repoName repoPRs - Patch Harvester image after merging PRs or checkout branches, always build image *

  Example: pr2pt harvester 0

* Authorized users only