Kubernetes Redis with High Availability
Switch branches/tags
Clone or download
Latest commit b84f2f6 May 15, 2017
Permalink
Failed to load latest commit information.
example Bump to 3.0.1 May 15, 2017
images Write documents #15 May 14, 2017
test Write documents #15 May 14, 2017
.gitignore Fix the commented part #3 Apr 23, 2017
.python-version Write test template #3 Apr 23, 2017
CHANGELOG.md Bump to 3.0.1 May 15, 2017
LICENSE Clarify license Mar 17, 2017
README.md Fix documentation #17 May 15, 2017

README.md

Kubernetes Redis with High Availability

Redis Images and Samples in this project are implemented using the latest features of Kubernetes:

  • StatefulSet
  • Init Container

Requirements

  • Kubernetes 1.6 cluster
  • Redis 3.2

Quick Start

If you already have a Kubernetes cluster, you can deploy High Availability Redis using the following command:

$ kubectl create -f example/
service "redis-sentinel" created
statefulset "redis-sentinel" created
service "redis-server" created
statefulset "redis-server" created
pod "console" created

Accessing Redis

You can access Redis server using console pod:

$ kubectl exec -ti console -- /bin/bash
root@console:# export MASTER_IP="$(redis-cli -h redis-sentinel -p 26379 sentinel get-master-addr-by-name mymaster | head -1)"
root@console:# export SERVER_PASS="_redis-server._tcp.redis-server.default.svc.cluster.local"
root@console:# redis-cli -h "${MASTER_IP}" -a "${SERVER_PASS}" set foo bar
OK
root@console:# redis-cli -h redis-server -a "${SERVER_PASS}" get foo
bar

The passowrd is necessary for preventing independent Redis clusters from merging together in some circumstances.

Scale Up and Down

With tarosky/k8s-redis-ha, you can scale up/down Redis servers and Redis sentinels like the normal Deployment resources:

$ kubectl scale --replicas=5 statefulset/redis-sentinel
statefulset "redis-sentinel" scaled
$ kubectl scale --replicas=5 statefulset/redis-server
statefulset "redis-server" scaled

After these scale up/down, the expected number of available slaves in the Redis set are reset automatically.

Sample Code in Python

$ kubectl exec -ti console -- ipython
In [1]: from redis import StrictRedis

In [2]: from redis.sentinel import Sentinel

In [3]: sentinel = Sentinel([
   ...:         ('redis-sentinel-0.redis-sentinel.default.svc.cluster.local', 26379),
   ...:         ('redis-sentinel-1.redis-sentinel.default.svc.cluster.local', 26379),
   ...:         ('redis-sentinel-2.redis-sentinel.default.svc.cluster.local', 26379)
   ...:     ], socket_timeout=0.1
   ...: )

In [4]: master = sentinel.master_for(
   ...:     'mymaster',
   ...:     password='_redis-server._tcp.redis-server.default.svc.cluster.local',
   ...:     socket_timeout=0.1
   ...: )

In [5]: slave = sentinel.slave_for(
   ...:     'mymaster',
   ...:     password='_redis-server._tcp.redis-server.default.svc.cluster.local',
   ...:     socket_timeout=0.1
   ...: )

In [6]: master.set('foo', 'bar')
Out[6]: True

In [7]: slave.get('foo')
Out[7]: b'bar'

Running the Test Script

$ pyvenv .venv
$ source .venv/bin/activate
$ pip install -r test/requirements.txt

You can run the test command using the following command:

$ KUBE_NAMESPACE='{{Your name space}}' nosetests test/test.py