Mirrors/flux2
1
0
Fork 0
mirror of synced 2026-02-06 10:55:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add workflow overview to get started guide

Browse Source
This commit is contained in:
stefanprodan
2020-06-22 10:02:17 +03:00
parent c0a8604f90
commit 041b475f49
1 changed files with 57 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

92
docs/get-started/index.md
View File

@@ -2,11 +2,7 @@
## Prerequisites
* Kubernetes >= 1.14
* kubectl >= 1.18
* git
You will need to have Kubernetes set up.
You will need a Kubernetes cluster version 1.14 or newer and kubectl version 1.18.
For a quick local test, you can use `minikube`, `kubeadm` or `kind`.
Any other Kubernetes setup will work as well though.
@@ -14,6 +10,13 @@ In order to follow the guide you'll need a GitHub account and a
[personal access token](https://help.github.com/en/github/authenticating-to-github/creating-a-personal-access-token-for-the-command-line)
that can create repositories.
Export your GitHub personal access token and username:
```sh
export GITHUB_TOKEN=<your-token>
export GITHUB_USER=<your-username>
```
## Install the toolkit CLI
To install the latest `tk` release run:
@@ -33,7 +36,20 @@ To configure your shell to load tk completions add to your bash profile:
. <(tk completion)
```
Verify that your cluster satisfies the prerequisites with:
## GitOps workflow
You'll be using a dedicated Git repository e.g. `fleet-infra` to manage one or more Kubernetes clusters.
This guide assumes that you have two clusters, one for staging and one for production.
Using the toolkit CLI you'll do the following:
- configure each cluster to synchronise with a directory inside the fleet repository
- register app sources (git repositories) that contain plain Kubernetes manifests or Kustomize overlays
- configure app deployments on both clusters (pre-releases on staging, semver releases on production)
## Staging bootstrap
Verify that your staging cluster satisfies the prerequisites with:
```text
$ tk check --pre
@@ -44,29 +60,20 @@ $ tk check --pre
✔ prerequisites checks passed
```
## Bootstrap
You'll be using a dedicated Git repository e.g. `fleet-infra` to manage one or more Kubernetes clusters.
First export your GitHub personal access token and GitHub username:
```sh
export GITHUB_TOKEN=<your-token>
export GITHUB_USER=<your-username>
```
The bootstrap command creates a repository if one doesn't exist and
commits the toolkit components manifests to the master branch at the specified path.
Then it configures the target cluster to synchronize with the specified path inside the repository.
Run the bootstrap command:
```sh
tk bootstrap github \
--owner=$GITHUB_USER \
--repository=fleet-infra \
--path=dev-cluster \
--path=staging-cluster \
--personal
```
The bootstrap command creates a repository if one doesn't exist and
commits the toolkit components manifests to the master branch at the specified path.
Then it configures the target cluster to synchronize with the specified path inside the repository.
If you wish to create the repository under a GitHub organization:
```sh
@@ -75,13 +82,13 @@ tk bootstrap github \
--repository=<repo-name> \
--team=<team1-slug> \
--team=<team2-slug> \
--path=dev-cluster
--path=staging-cluster
```
Example output:
```text
$ tk bootstrap github --owner=gitopsrun --repository=fleet-infra --path=dev-cluster --team=devs
$ tk bootstrap github --owner=gitopsrun --repository=fleet-infra --path=staging-cluster --team=devs
► connecting to github.com
✔ repository created
@@ -118,11 +125,12 @@ deployment "kustomize-controller" successfully rolled out
If you prefer GitLab, export `GITLAB_TOKEN` env var and use the command [tk bootstrap gitlab](../cmd/tk_bootstrap_gitlab.md).
It is safe to run the bootstrap command as many times as you want.
If the toolkit components are present on the cluster,
the bootstrap command will perform an upgrade if needed.
!!! hint
It is safe to run the bootstrap command as many times as you want.
If the toolkit components are present on the cluster,
the bootstrap command will perform an upgrade if needed.
## Create a GitOps workflow
## Staging workflow
Clone the repository with:
@@ -131,14 +139,14 @@ git clone https://github.com/$GITHUB_USER/fleet-infra
cd fleet-infra
```
Create a git source pointing to a public repository:
Create a git source pointing to a public repository master branch:
```sh
tk create source git webapp \
--url=https://github.com/stefanprodan/podinfo \
--branch=master \
--interval=30s \
--export > ./dev-cluster/webapp-source.yaml
--export > ./staging-cluster/webapp-source.yaml
```
Create a kustomization for synchronizing the common manifests on the cluster:
@@ -150,7 +158,7 @@ tk create kustomization webapp-common \
--prune=true \
--validate=client \
--interval=1h \
--export > ./dev-cluster/webapp-common.yaml
--export > ./staging-cluster/webapp-common.yaml
```
Create a kustomization for the backend service that depends on common:
@@ -165,7 +173,7 @@ tk create kustomization webapp-backend \
--interval=10m \
--health-check="Deployment/backend.webapp" \
--health-check-timeout=2m \
--export > ./dev-cluster/webapp-backend.yaml
--export > ./staging-cluster/webapp-backend.yaml
```
Create a kustomization for the frontend service that depends on backend:
@@ -180,13 +188,13 @@ tk create kustomization webapp-frontend \
--interval=10m \
--health-check="Deployment/frontend.webapp" \
--health-check-timeout=2m \
--export > ./dev-cluster/webapp-frontend.yaml
--export > ./staging-cluster/webapp-frontend.yaml
```
Push changes to origin:
```sh
git add -A && git commit -m "add dev webapp" && git push
git add -A && git commit -m "add staging webapp" && git push
```
In about 30s the synchronization should start:
@@ -214,7 +222,20 @@ service/backend ClusterIP 10.52.10.22 <none> 9898/TCP,9999/TCP 4
service/frontend ClusterIP 10.52.9.85 <none> 80/TCP 3m31s
```
## Production workflow
!!! note
From this moment forward, any changes made to the webapp
master branch will be synchronised with the cluster.
If a Kubernetes manifest is removed from source, the reconclier will remove it from your cluster. If you
delete a kustomization from the fleet infra repo, the reconciler will remove all Kubernetes objects that
were previously applied from that kustomization.
If you alter the webapp deployment using `kubectl edit`, the changes will be reverted to match
the state described in git. When dealing with an incident, you can pause the recitation of a
kustomization with `tk suspend kustomization <name>`. Once the debugging session
is over, you can re-enable the reconciliation with `tk resume kustomization <name>`.
## Production bootstrap
On production clusters, you may wish to deploy stable releases of an application.
When creating a git source instead of a branch, you can specify a git tag or a semver expression.
@@ -235,6 +256,8 @@ Pull the changes locally:
git pull
```
## Production workflow
Create a git source using a semver range to target stable releases:
```sh
@@ -325,4 +348,3 @@ $ watch tk get kustomizations
✔ gitops-system last applied revision master/d751ea264d48bf0db8b588d1d08184834ac8fec9
✔ webapp last applied revision 4.0.5/f43f9b2eb6766e07f318d266a99d2ec7c940b0cf
```