Skip to content

Latest commit

 

History

History
48 lines (35 loc) · 1.67 KB

quick_start.md

File metadata and controls

48 lines (35 loc) · 1.67 KB
Raw

Quick Start

Compilation and Startup of KubeBrain

Badger-based Stand-alone Storage

KubeBrain adapts to dgraph-io/badger to implement a stand-alone version for testing.

make badger
./bin/kube-brain --key-prefix "/"

TiKV-based High-availability Storage

KubeBrain adapts to TiKV to implement kubernetes meta-information storage based on distributed key-value database.

TiKV Deployment

For TiKV cluster deployment, please refer to TiUP. For the local test environment, you can use tiup playground.

tiup playground --mode tikv-slim --kv 1 --pd 1

Compilation and Startup

make tikv
./bin/kube-brain --key-prefix "/" --pd-addr=127.0.0.1:2379 --port=3379 --peer-port=3380 --compatible-with-etcd=true

When multiple KubeBrain share a tikv cluster

  • key-prefix is consistent with the etcd-prefixparameter of the apiserver corresponding to the apiserver.
  • compatible-with-etcd should set true to make all nodes provide all etcd APIs used in apiserver

APIServer

Use the Community Version of APIServer

The community version of the APIServer storage only supports the API of etcd3, and KubeBrain is compatible with the API of etcd3 used in the community version of APIServer, but this method does not take full advantage of KubeBrain.

./kube-apiserver  /
  # ...
  --storage-backend=etcd3 \
  --etcd-servers=http://127.0.0.1:3379 \
  # ...

Using the Custom APIServer

The customized APIServer accesses the storage system by extending the storage interface of APIServer, and integrating functions such as cluster state checking, read-write separation of load, etc. in the SDK. The custom api-server will be released soon.