Skip to content

Latest commit

 

History

History
56 lines (39 loc) · 2.12 KB

kubefork-installation-en.md

File metadata and controls

56 lines (39 loc) · 2.12 KB

How to deploy kubefork

This document describes how to deploy kubefork in a cluster.

Requirement

Go 1.16+ kubernetes 1.25+

In addition, please install the following into your cluster in advance deployment-duplicator Istio 1.15+ Emissary-ingress 3.1+

Introduction of kubefork-controller

See kubefork-controller.

Introduction of ForkManager

In order to perform kubefork, it is necessary to apply the ForkManager resource, which configures which headers are looked at and which hosts and services correspond to. The most basic ForkManager manifest is as follows

apiVersion: fork.k8s.wantedly.com/v1beta1
kind: ForkManager
metadata:
  name: kubefork-manager
  namespace: manager
spec:
  ambassadorID: fork-ambassador
  headerKey: x-fork-identifier
  upstreams:
  - host: example.com
    original: your-service:80

In this case, the header key referenced to distribute communications is x-fork-identifier, and communications to example.com are normally sent to port 80 of the service your-service.

For a detailed specification of ForkManager, see forkmanager_types.go.

Installation of kubeforkctl

After applying ForkManager, you should install the command line tool kubeforkctl.

References

Currently, in order to do kubefork, users must have the ability to propagate certain headers in their applications. See the kubefork documentation for details.