Skip to content

Latest commit

 

History

History
406 lines (349 loc) · 12.5 KB

README.md

File metadata and controls

406 lines (349 loc) · 12.5 KB
Raw

REDIS

This repo is the result of studing Redis. See references section to understand better the path follow.

Definition

Redis means REmote DIctionary Server.

  • All commands in Redis are executed atomically. This is guaranteed for the way to Redis works.

Install

To install redis, we can go to the official site: https://redis.io/download

Or we can start them in docker:

$ docker run --name local-redis -p 6379:6379 -d redis

To run CLI we can do:

$ docker exec -it local-redis redis-cli

CLI basic commands

To create a key, we can execute:

127.0.0.1:6379> SET "last_user" "iundarigun"
OK

To get a key, we can execute:

127.0.0.1:6379> GET "last_user"
"iundarigun"

We can append string into an existing key:

127.0.0.1:6379> APPEND "last_user" " additional"

To delete a key, we can execute:

127.0.0.1:6379> DEL "last_user"
(integer) 1

List all keys:

127.0.0.1:6379> KEYS *
1) "last_user"

It is a good pattern structure your key with items that make sanse for you and separate by :. Setting a key structure:

127.0.0.1:6379> SET "result:2019-10-29:test-coverage" "100%"
OK

Setting more than key at the same time:

127.0.0.1:6379> MSET "result:2019-10-28:test-coverage" "98%" "result:2019-10-27:test-coverage" "93%" "result:2019-10-26:test-coverage" "95%"
OK

Get keys from results of 2019:

127.0.0.1:6379> KEYS "result:2019-*"

Get keys from results of month 10, 11, or 12 from 2019:

127.0.0.1:6379> KEYS "result:2019-1?-*"

To search with some conditions at the same time. Get keys from results of month 10 and 11 from 2019, but not 12:

127.0.0.1:6379> KEYS "result:2019-1[01]-*"

Saving more info in the same key, using a set of values:

127.0.0.1:6379> HSET "result:2019-11-01:test-coverage" "coverage" "97%"
127.0.0.1:6379> HSET "result:2019-11-01:test-coverage" "exclude package" "br.com.devcave.*.domain.*"

To set some values at the same time:

127.0.0.1:6379> HMSET "result:2019-11-02:test-coverage" "coverage" "94%" "exclude package" "br.com.devcave.*.domain.*"

To get an especific value:

127.0.0.1:6379> HGET "result:2019-11-01:test-coverage" "exclude package"

To get all values of a set in a key:

127.0.0.1:6379> HGETALL "result:2019-11-01:test-coverage"

To get all values without key:

127.0.0.1:6379> HVALS "result:2019-11-01:test-coverage"

To remove an element of a set:

127.0.0.1:6379> HDEL "result:2019-11-01:test-coverage" "exclude package"

Using TTL

Setting a ttl. Expire is in seconds:

127.0.0.1:6379> HMSET "session:user:2110" "name" "Oriol" "last name" "Canalias"
127.0.0.1:6379> EXPIRE "session:user:2110" 60

In some instructions, we can do the same on a single instruction:

127.0.0.1:6379> SET "some:words:here" "Wellcome with expiration" EX 20

To view the expire time left:

127.0.0.1:6379> TTL "session:user:2110"

Using NX and XX

When set a key, if we use NX modifier, the valeu is only set if key doesn't exists:

127.0.0.1:6379> SET "session:user:860" "iundarigun" NX

When set a key, if we use XX modifier, the valeu is only set if key exists:

127.0.0.1:6379> SET "session:user:860" "iundarigun" XX

Using to count

We can use a value to count or some a value. If the key is not exists, redis create it:

  • Increment (unit): 
    127.0.0.1:6379> INCR page:/home/2019-10-28
  • Increment with value: 
    127.0.0.1:6379> INCRBY page:/home/2019-10-28 10
127.0.0.1:6379> INCRBYFLOAT page:/home/2019-10-28 10.5
  • Decrement: 
    127.0.0.1:6379> DECR page:/home/2019-10-28
  • Decrement with value: 
    127.0.0.1:6379> DECRBY page:/home/2019-10-28 10
127.0.0.1:6379> DECRBYFLOAT page:/home/2019-10-28 10.5

Saving bits

To save bits, we can save a group of bits (true/false) for a especific key. The function is similar to HSET/HGET, but the id of the set (not confunding with the key of redis) is always a number

127.0.0.1:6379> SETBIT enter:2019-10-28 9945 1

To get:

127.0.0.1:6379> GETBIT enter:2019-10-28 9945

To count bits(number of trues) in a set:

127.0.0.1:6379> BITCOUNT enter:2019-10-28

Repeat true ids on some keys, need some and store the result in another key:

127.0.0.1:6379> BITOP AND enter:resume enter:2019-10-30 enter:2019-10-31

Count if existes some true ids on some keys, need some and store the result in another key:

127.0.0.1:6379> BITOP OR enter:resume enter:2019-10-30 enter:2019-10-31

Working with lists

To create a list with the last elements of some list, like a list of the last news, we can use LPUSH to put elements in a list. If we can to delimited the list to a finit number of elements we can use the LTRIM instruction. Other instructions that you can use are LINDEX to get some element passing position and LLEN to get the size of the list.

127.0.0.1:6379> LPUSH "last_news" "new of 2019-10-30 10:00"
...
127.0.0.1:6379> LINDEX "last_news" 0
127.0.0.1:6379> LLEN "last_news"
...
127.0.0.1:6379> LTRIM "last_news" 0 2

We can get some elements in a range:

127.0.0.1:6379> LRANGE "last_news" 1 2

If we need to put some elements at the end of the list, we can use RPUSH

127.0.0.1:6379> RPUSH "last_news" "new of 2019-10-30 09:00"

Using list like a queue

We can use this kindle of list like a queue, using RPUSH. To get the elements to the queue "deleting" the element, we can use LPOP:

127.0.0.1:6379> RPUSH "queue:email-confirmation" "user1@devcave.com.br"
127.0.0.1:6379> LPOP "queue:email-confirmation"

To implement a LIFO, we can use RPOP instruction.

If the queue is empty, we can block the pop for a time, expressed in seconds(0 to wait forever):

127.0.0.1:6379> BLPOP "queue:email-confirmation" 30 
127.0.0.1:6379> BRPOP "queue:email-confirmation" 30

Collections

A collection has a list of unic elements inside. To add elements:

127.0.0.1:6379> SADD "friends:iundarigun" "friend1" "friend2"

Counting elements:

127.0.0.1:6379> SCARD "friends:iundarigun"

Showing elements:

127.0.0.1:6379> SMEMBERS "friends:iundarigun"

Validating existing elements:

127.0.0.1:6379> SISMEMBER "friends:iundarigun" "friend2"

Removing:

127.0.0.1:6379> SREM "friends:iundarigun" "friend1"

If we want to get a intersection of two collections:

127.0.0.1:6379> SINTER "friends:iundarigun" "friends:iunda"

To get differences:

127.0.0.1:6379> SDIFF "friends:iundarigun" "friends:iunda"

Union of two collections without repetions:

127.0.0.1:6379> SUNION "friends:iundarigun" "friends:iunda"

Making a ranking

To create a ranking, we can use a hash key to declare the players:

127.0.0.1:6379> HSET "players:id:991" "username" "iundarigun" "points" 12

If we want to get all informations, we can use HGETALL, like we see before. If we want increase the points of the player, we can use some instructions, HINCRBY or HINCRBYFLOAT. Doesn't exists HINCRBY!

127.0.0.1:6379> HINCRBY "players:id:991" "points" 3

To create an order list, we can use zset (Sorted Set). To add values in a key, we use ZADD with key, value used to order and identifyer of the element:

127.0.0.1:6379> ZADD "players:ranking" 15 "player1"

We count the elements using:

127.0.0.1:6379> ZCARD "players:ranking"

We can get the firsts members by order using (asc):

127.0.0.1:6379> ZRANGE "players:ranking" 0 3 WITHSCORES

We can get descending order:

127.0.0.1:6379> ZREVRANGE "players:ranking" 0 3 WITHSCORES

To get all elements, we can use negative positions:

127.0.0.1:6379> ZREVRANGE "players:ranking" 0 -1 WITHSCORES

To get score from some elements, or the positions, we can use:

127.0.0.1:6379> ZSCORE "players:ranking" "player1"
127.0.0.1:6379> ZRANK "players:ranking" "player1"
127.0.0.1:6379> ZREVRANK "players:ranking" "player1"

To increase points:

127.0.0.1:6379> ZINCRBY "players:ranking" 5000 "player1"

Transactions

We can do transactions using the key words multi and exec

127.0.0.1:6379> multi
OK
127.0.0.1:6379> set num 10
QUEUED
127.0.0.1:6379> incrby num 10
QUEUED
127.0.0.1:6379> exec
1) OK
2) (integer) 20

To cancel transaction we can use the key word discard:

127.0.0.1:6379> multi
OK
127.0.0.1:6379> set num 10
QUEUED
127.0.0.1:6379> incrby num 10
QUEUED
127.0.0.1:6379> discard
OK

If we want cancel the transaction if someone changes some key:

127.0.0.1:6379> watch num
OK
127.0.0.1:6379> multi
OK
127.0.0.1:6379> set num 10
QUEUED

if until do exec in this transaction, someone change the key, the exec will not work:

127.0.0.1:6379> exec
(nil)

Pub-Sub

Redis is used like a queue. The operation of PubSub seems very simple, but this is one of the main characteristics, like its performance

To put message in a channel:

127.0.0.1:6379> PUBLISH "channel-1" "message 1"

If the return is 0 means that doesn't exists any subscriber at this channel and no one receives the message. To create a subscribe:

127.0.0.1:6379> SUBSCRIBE "channel-1"

Lua Scripts

Sometimes we need to execute some instructions without any interaction with data from others connections. Redis garantee atomic executions from one instrucions, but by default no for more than one. Lua Script resolve this problem using EVAL instruction:

127.0.0.1:6379> EVAL "local members  = redis.call('smembers', KEYS[1]) \n for mem=1,#members do return members[mem] end" 1 "somesetkey"

The first argument is the Lua script, the second is the size of the array of keys, and the next arguments are the list of keys.

Security

We can secure our redis configuring a password. We can do this in the configfile or on cli:

127.0.0.1:6379> config set requirepass root

After this we need auth before execute any command:

127.0.0.1:6379> auth root

Cloud redis

app.redislabs.com is a cloud redis. We can create an account to play.

Instance configuration

Storage

We can configure two ways to store data:

  • RDB is the most popular. It is the default option. Data is persisted asynchronous, on scheduled intervals, that we can configure. In the case of server crashes, we can lose some data.
  • AOF configuration saves a log for every operation, saved in a single file. When Redis restarts data will be rebuilding from this log file. No data will be lost, but performance can be reduced.

Database schema

Redis has this concept, but by default the first database schema is active and we change the active database schema we can use:

127.0.0.1:6379> select <number of schema>

By default, Redis brings to us 16 database schemas, identified for a number, and the zero is the default database's schema.

Cluster and Sentinel

Sentinel is a way to brings high availability and recover failures, because we can configure some slaves to get the place of master when something wrong happens with the master.

Cluster brings high availability too, but not only this, but also brings data sharing, using the hash of key to decided the slot destination. Clusters divide 16.000 slots aprox, and the destination slot is decided using the module of hash to the number of slots.

Spring and Redis

We can use redis for a few thinks with Spring-boot. In this repo, there is a folder redis with Kotlin and Spring boot application with the next examples:

  • Use RedisTemplate to get/set values - ProgrammerRepositoryImpl.kt
  • Use Spring-Data to manage crud operations - StudentRepository.kt
  • Using as a cache - Sample in CacheService.kt
  • Tests with embedded server - RedisApplicationTests.kt

References