polukovy/hsa-redis-cluster-with-probability-cache

Redis cluster with probability cache

redis-cluster

Docker Compose template of Redis cluster

The template defines the topology of the Redis cluster

master:
  image: redis:3
slave:
  image: redis:3
  command: redis-server --slaveof redis-master 6379
  links:
    - master:redis-master
sentinel:
  build: sentinel
  environment:
    - SENTINEL_DOWN_AFTER=5000
    - SENTINEL_FAILOVER=5000    
  links:
    - master:redis-master
    - slave

There are following services in the cluster,

• master: Redis master
• slave: Redis slave
• sentinel: Redis sentinel

The sentinels are configured with a "mymaster" instance with the following properties -

sentinel monitor mymaster redis-master 6379 2
sentinel down-after-milliseconds mymaster 5000
sentinel parallel-syncs mymaster 1
sentinel failover-timeout mymaster 5000

The details could be found in sentinel/sentinel.conf

The default values of the environment variables for Sentinel are as following

• SENTINEL_QUORUM: 2
• SENTINEL_DOWN_AFTER: 30000
• SENTINEL_FAILOVER: 180000

Play with it

Build the sentinel Docker image

docker-compose build

Start the redis cluster

docker-compose up -d

Check the status of redis cluster

docker-compose ps

The result is

         Name                        Command               State          Ports        
--------------------------------------------------------------------------------------
redis_master_1     docker-entrypoint.sh redis ...   Up      6379/tcp            
redis_sentinel_1   docker-entrypoint.sh redis ...   Up      26379/tcp, 6379/tcp 
redis_slave_1      docker-entrypoint.sh redis ...   Up      6379/tcp     

Scale out the instance number of sentinel

docker-compose scale sentinel=3

Scale out the instance number of slaves

docker-compose scale slave=2

Check the status of redis cluster

docker-compose ps

The result is

         Name                        Command               State          Ports        
--------------------------------------------------------------------------------------
redis_master_1     docker-entrypoint.sh redis ...   Up      6379/tcp            
redis_sentinel_1   docker-entrypoint.sh redis ...   Up      26379/tcp, 6379/tcp 
redis_sentinel_2   docker-entrypoint.sh redis ...   Up      26379/tcp, 6379/tcp 
redis_sentinel_3   docker-entrypoint.sh redis ...   Up      26379/tcp, 6379/tcp 
redis_slave_1      docker-entrypoint.sh redis ...   Up      6379/tcp            
redis_slave_2      docker-entrypoint.sh redis ...   Up      6379/tcp            

Execute the test scripts

./test.sh

to simulate stop and recover the Redis master. And you will see the master is switched to slave automatically.

Or, you can do the test manually to pause/unpause redis server through

docker pause rediscluster_master_1
docker unpause rediscluster_master_1

And get the sentinel information with the following commands

docker exec rediscluster_sentinel_1 redis-cli -p 26379 SENTINEL get-master-addr-by-name mymaster

Reference useful documentation

• https://redis.io/topics/sentinel
• https://redis.io/commands/hset

Application logic

• Create data
  • logic save data to redis and create a scheduler with ttl less than redis TTL in this example (Notes: The logic of how we create a scheduler not optimal better to use ScheduledExecutorService better thread reusability.) Redis TTL 120 Refresh scheduler TTL 105

curl -X POST \
  'http://127.0.0.1:9001/keys' \
  -H 'Content-Type: application/json' \
  -H 'Postman-Token: f6d28274-a901-437e-ad4c-93c0143d6e2d' \
  -H 'cache-control: no-cache' \
  -d '{
    "id": "user"
}'

• Application logs

2021-02-16 11:51:38.207  INFO 45706 --- [nio-9001-exec-1] com.levik.redis.service.UserService      : saveOrUpdate -> User(id=user)
2021-02-16 11:51:38.698  INFO 45706 --- [nio-9001-exec-1] c.l.r.service.RefreshCacheUserService    : Start scheduler for User(id=user) close ttl 105000 millisecond now 2021-02-16T11:51:38.695703

2021-02-16 11:53:23.703  INFO 45706 --- [efresh-key-user] c.l.r.service.RefreshCacheUserService    : Trigger update -> User(id=user) now 2021-02-16T11:53:23.703684
2021-02-16 11:53:23.704  INFO 45706 --- [efresh-key-user] com.levik.redis.service.UserService      : saveOrUpdate -> User(id=user)
2021-02-16 11:53:23.724  INFO 45706 --- [efresh-key-user] c.l.r.service.RefreshCacheUserService    : Start scheduler for User(id=user) close ttl 105000 millisecond now 2021-02-16T11:53:23.724136

• Get data by id

curl -X GET 'http://127.0.0.1:9001/keys?id=user'

• Application logs

2021-02-16 11:52:07.211  INFO 45706 --- [nio-9001-exec-4] com.levik.redis.service.UserService      : getUserById -> user