redis安装

简单介绍redis以及ubuntu和windows下如何安装redis和配置文件详解

07net01.com 发布于 2015-08-22 01:48:02
分类：IT技术
阅读(57)
评论

机缘巧合决定了解一下redis。
首先看个关于数据库的图，让想学数据库的同学有一个整体的概念。



好勒，下面说redis。

了解redis

Redis是一个开源，先进的key-value存储，并用于构建高性能，可扩展的Web应用程序的完美解决方案。 

一句话概括就是：开源的/基于C语言的/基于内存亦可持久化的高性能的NoSql的键值对数据库。
redis存储结构丰富：字符串类型、散列类型、列表类型、集合类型、有序集合类型。
内存存储与持久化：1秒写入10万键值对、异步写入硬盘
功能丰富：用作数据库、缓存系统、队列
简单稳定：使用命令来读取数据，redis命令比sql语言简单的多 

比如我们在post表中查询一个id 为1的学生SQL语句这写：

select title from post where id=1 ;

redis语句只需要一句即可：

HGET post:1 title

了解完毕我们来安装一下，很简单。

首先ubuntu下安装redis

1、打开终端，输入一下命令

$sudo apt-get update $sudo apt-get install
redis-server

确认安装。等待安装完成。
2、启动redis。

$redis-server

3、检查Redis是否在工作

$redis-cli

上面的提示127.0.0.1是本机的IP地址，6379为Redis服务器运行的端口。现在输入PING命令，回复pong这说明你已经成功地安装Redis在您的机器上。

Windows下安装redis

1、下载地址：https://github.com/dmajkic/redis/downloads。下载windows版本的redis包。最新的是双版本，32、64位都在一起。
2、下载后解压到自己的某个文件夹。我选择64位的并把让它放在：E:\redis\64bit。在文件夹中可执行的2进制文件共有5个，分别代表如下： 



3、打开一个cmd窗口 使用cd命令切换目录到E:\redis\64bit 运行 redis-server.exe redis.conf 



4、出现上面那个图就表示启动服务成功了，这时候别启一个cmd窗口，原来的不要关闭，不然就无法访问服务端了。
5、切换到redis目录下运行 redis-cli.exe -h 127.0.0.1 -p 6379 出现下图： 



我们可以输入一个简单的测试数据测试。 

当我们访问服务端时，我们观察刚刚开启的服务端能看到如下数据。 



这时候，就已经完成配置了，现在说下它的的redis.conf配置文件。我摘抄了一个详细的说明，留博客备查

下面是配置相关项的说明

# Redis configuration file example
# Note on units: when memory size is needed, it is possible to specifiy # it in the usual form of 1k 5GB 4M and so
forth: # # 1k => 1000 bytes # 1kb => 1024 bytes # 1m => 1000000 bytes # 1mb => 1024*1024 bytes
# 1g => 1000000000 bytes # 1gb => 1024*1024*1024 bytes # # units are case insensitive
so 1GB 1Gb 1gB are all the same. # By default Redis does not run as a daemon. Use 'yes' if you
need it. # Note that Redis will write a pid file in /var/run/redis.pid when daemonized. daemonize no Redis默认不是以守护进程的方式运行，可以通过该配置项修改，使用yes启用守护进程 # When running daemonized,
Redis writes a pid file in /var/run/redis.pid by # default. You can specify a custom pid file location here. pidfile /var/run/redis.pid 当Redis以守护进程方式运行时，Redis默认会把pid写入/var/run/redis.pid文件，可以通过pidfile指定
# Accept connections on the specified port, default is 6379. # If port 0 is specified
Redis will not listen on a TCP socket. port 6379 指定Redis监听端口，默认端口为6379 # If you want you can bind a single interface, if the
bind option is not # specified all the interfaces will listen for incoming connections. # # bind 127.0.0.1 绑定的主机地址
# Specify the path for the unix socket that will be used to listen for # incoming connections. There is no default, so Redis will not listen
# on a unix socket when not specified. # # unixsocket /tmp/redis.sock # unixsocketperm 755 # Close the connection after a client is idle for N
seconds (0 to disable) timeout 0 当 客户端闲置多长时间后关闭连接，如果指定为0，表示关闭该功能
# Set server verbosity to 'debug' # it can be one of: # debug (a lot of information, useful for development/testing) # verbose (many rarely useful
info, but not a mess like the debug level) # notice (moderately verbose, what you want in production probably) # warning (only very important / critical messages
are logged) loglevel verbose 指定日志记录级别，Redis总共支持四个级别：debug、verbose、notice、warning，默认为verbose # Specify the log file name. Also 'stdout' can be used to force # Redis to log on the
standard output. Note that if you use standard # output for logging but daemonize, logs will be sent to /dev/null logfile stdout 日志记录方式，默认为标准输出，如果配置Redis为守护进程方式运行，而这里又配置为日志记录方式为标准输出，则日志将会发送给/dev/null # To enable
logging to the system logger, just set 'syslog-enabled' to yes,
# and optionally update the other syslog parameters to suit your needs. # syslog-enabled no # Specify the syslog identity. # syslog-ident redis # Specify the syslog facility. Must be USER or between
LOCAL0-LOCAL7. # syslog-facility local0 # Set the number of databases.
The default database is DB 0, you can select # a different one on a per-connection basis using SELECT <dbid> where # dbid is a
number between 0 and 'databases'-1 databases 16 设置数据库的数量，默认数据库为0，可以使用SELECT <dbid>命令在连接上指定数据库id ################################
SNAPSHOTTING ################################# # # Save the DB on disk: # # save <seconds> <changes> # # Will save the DB if both the given number of seconds and the given
# number of write operations against the DB occurred. # # In the example below the behaviour will be to save: # after 900 sec (15 min) if at
least 1 key changed # after 300 sec (5 min) if at least 10 keys changed # after 60 sec if at
least 10000 keys changed # # Note: you can disable saving at all commenting all the "save" lines. save 900 1 save 300 10 save 60 10000 分别表示900秒（15分钟）内有1个更改，300秒（5分钟）内有10个更改以及60秒内有10000个更改。
指定在多长时间内，有多少次更新操作，就将数据同步到数据文件，可以多个条件配合 # Compress string objects using LZF when dump .rdb databases? # For default that's set to 'yes'
as it's almost always a win. # If you want to save some CPU in the saving child set it to 'no' but # the dataset will likely
be bigger if you have compressible values or keys. rdbcompression yes 指定存储至本地数据库时是否压缩数据，默认为yes，Redis采用LZF压缩，如果为了节省CPU时间，可以关闭该选项，但会导致数据库文件变的巨大 # The filename where to dump the DB dbfilename dump.rdb
指定本地数据库文件名，默认值为dump.rdb # The working directory. # # The DB will be written inside this directory, with the filename specified # above using the 'dbfilename' configuration directive. # # Also
the Append Only File will be created inside this directory. # # Note that you must specify a directory here, not a file name. dir ./ 指定本地数据库存放目录 ################################# REPLICATION
################################# # Master-Slave replication. Use slaveof to make a Redis instance a copy of # another Redis server. Note that the configuration is local to the
slave # so for example it is possible to configure the slave to save the DB with a # different interval, or to listen to another port, and so on.
# # slaveof <masterip> <masterport> slaveof <masterip> <masterport> 设置当本机为slav服务时，设置master服务的IP地址及端口，在Redis启动时，它会自动从master进行数据同步 # If the master is password protected (using the "requirepass" configuration #
directive below) it is possible to tell the slave to authenticate before # starting the replication synchronization process, otherwise the master will # refuse the slave
request. # # masterauth <master-password> masterauth <master-password> 当master服务设置了密码保护时，slav服务连接master的密码 # When a slave lost the connection with the master, or when the
replication # is still in progress, the slave can act in two different ways: # # 1) if slave-serve-stale-data is set to 'yes'
(the default) the slave will # still reply to client requests, possibly with out of data data, or the # data set may just be
empty if this is the first synchronization. # # 2) if slave-serve-stale data is set to 'no' the slave
will reply with # an error "SYNC with master in progress" to all the kind of commands
# but to INFO and SLAVEOF. # slave-serve-stale-data yes # Slaves send PINGs to server in a predefined interval. It's possible to change
# this interval with the repl_ping_slave_period option. The default value is 10 # seconds. # # repl-ping-slave-period 10 # The following option sets
a timeout for both Bulk transfer I/O timeout and # master data or ping response timeout. The default value is 60 seconds. #
# It is important to make sure that this value is greater than the value # specified for repl-ping-slave-period otherwise a timeout will be detected # every time there is low
traffic between the master and the slave. # # repl-timeout 60 ################################## SECURITY ################################### # Require clients to issue AUTH <PASSWORD> before
processing any other # commands. This might be useful in environments in which you do not trust # others with access to the
host running redis-server. # # This should stay commented out for backward compatibility and because most # people do not need auth (e.g. they run their own servers). #
# Warning: since Redis is pretty fast an outside user can try up to # 150k passwords per second against a good box. This means that you should # use a very strong password
otherwise it will be very easy to break. # # requirepass foobared requirepass foobared 设置Redis连接密码，如果配置了连接密码，客户端在连接Redis时需要通过AUTH <password>命令提供密码，默认关闭 # Command renaming. # # It is possilbe to change
the name of dangerous commands in a shared # environment. For instance the CONFIG command may be renamed into something # of hard to guess
so that it will be still available for internal-use # tools but not available for general clients. # # Example: # # rename-command CONFIG b840fc02d524045429941cc15f59e41cb7be6c52
# # It is also possilbe to completely kill a command renaming it into # an empty string: # # rename-command CONFIG "" ################################### LIMITS ####################################
# Set the max number of connected clients at the same time. By default there # is no limit, and it's up to the
number of file descriptors the Redis process # is able to open. The special value '0' means no limits.
# Once the limit is reached Redis will close all the new connections sending # an error 'max number of clients reached'. # # maxclients 128 maxclients 128 设置同一时间最大客户端连接数，默认无限制，Redis可以同时打开的客户端连接数为Redis进程可以打开的最大文件描述符数，如果设置
maxclients 0，表示不作限制。当客户端连接数到达限制时，Redis会关闭新的连接并向客户端返回max number of clients reached错误信息 # Don't use more memory than the specified amount of bytes.
# When the memory limit is reached Redis will try to remove keys with an # EXPIRE set. It will try to start freeing keys that are going to expire
# in little time and preserve keys with a longer time to live. # Redis will also try to remove objects
from free lists if possible. # # If all this fails, Redis will start to reply with errors to commands # that will use more
memory, like SET, LPUSH, and so on, and will continue # to reply to most read-only commands like GET. # # WARNING: maxmemory can be a good idea mainly if you
want to use Redis as a # 'state' server or cache, not as a real DB. When Redis is used as a real # database
the memory usage will grow over the weeks, it will be obvious if # it is going to use too much memory in the long run, and you'll have
the time # to upgrade. With maxmemory after the limit is reached you'll start to get # errors for write
operations, and this may even lead to DB inconsistency. # # maxmemory <bytes> maxmemory <bytes>指定Redis最大内存限制，Redis在启动时会把数据加载到内存中，达到最大内存后，Redis会先尝试清除已到期或即将到期的Key，当此方法处理 后，仍然到达最大内存设置，将无法再进行写入操作，但仍然可以进行读取操作。Redis新的vm机制，会把Key存放内存，Value会存放在swap区
# MAXMEMORY POLICY: how Redis will select what to remove when maxmemory # is reached? You can select among five behavior: # # volatile-lru -> remove
the key with an expire set using an LRU algorithm # allkeys-lru -> remove any key accordingly to the LRU algorithm # volatile-random -> remove a random key with an expire set # allkeys->random
-> remove a random key, any key # volatile-ttl -> remove the key with the nearest expire time (minor TTL) # noeviction -> don't expire at all, just return an
error on write operations # # Note: with all the kind of policies, Redis will return an error on write # operations, when there
are not suitable keys for eviction. # # At the date of writing this commands are: set setnx setex append # incr decr rpush lpush rpushx lpushx linsert lset rpoplpush sadd # sinter sinterstore
sunion sunionstore sdiff sdiffstore zadd zincrby # zunionstore zinterstore hset hsetnx hmset hincrby incrby decrby # getset mset msetnx exec sort # # The default is: # # maxmemory-policy volatile-lru # LRU and minimal
TTL algorithms are not precise algorithms but approximated # algorithms (in order to save memory), so you can select as well the sample # size to check. For instance for default Redis
will check three keys and # pick the one that was used less recently, you can change the sample size # using the following configuration directive. # # maxmemory-samples 3 ##############################
APPEND ONLY MODE ############################### # By default Redis asynchronously dumps the dataset on disk. If you can live # with the idea that the latest records will
be lost if something like a crash # happens this is the preferred way to run Redis. If instead you care a lot # about your data and don't want to that
a single record can get lost you should # enable the append only mode: when this mode is enabled Redis will append # every write operation received in the file appendonly.aof.
This file will # be read on startup in order to rebuild the full dataset in memory. # # Note that you can have both the async dumps and the
append only file if you # like (you have to comment the "save" statements above to disable the dumps). # Still if append only
mode is enabled Redis will load the data from the # log file at startup ignoring the dump.rdb file. # # IMPORTANT: Check the BGREWRITEAOF to check how to rewrite
the append # log file in background when it gets too big. appendonly no appendonly no指定是否在每次更新操作后进行日志记录，Redis在默认情况下是异步的把数据写入磁盘，如果不开启，可能会在断电时导致一段时间内的数据丢失。因为 redis本身同步数据文件是按上面save条件来同步的，所以有的数据会在一段时间内只存在于内存中。默认为no
# The name of the append only file (default: "appendonly.aof") # appendfilename appendonly.aof appendfilename appendonly.aof指定更新日志文件名，默认为appendonly.aof # The fsync() call
tells the Operating System to actually write data on disk # instead to wait for more data in the output buffer.
Some OS will really flush # data on disk, some other OS will just try to do it ASAP. # # Redis supports three different modes: # # no: don't fsync, just let the OS flush the data when it
wants. Faster. # always: fsync after every write to the append only log . Slow, Safest. # everysec: fsync only if one second passed since the last fsync. Compromise. # # The default is "everysec" that's usually
the right compromise between # speed and data safety. It's up to you to understand if you can relax this to # "no" that
will will let the operating system flush the output buffer when # it wants, for better performances (but if you can live with the idea of #
some data loss consider the default persistence mode that's snapshotting), # or on the contrary, use "always" that's very
slow but a bit safer than # everysec. # # If unsure, use "everysec". # appendfsync always appendfsync everysec # appendfsync no 指定更新日志条件，共有3个可选值：
no：表示等操作系统进行数据缓存同步到磁盘（快） always：表示每次更新操作后手动调用fsync()将数据写到磁盘（慢，安全）
everysec：表示每秒同步一次（折衷，默认值） # When the AOF fsync policy is set to always or everysec, and a background # saving process (a background
save or AOF log background rewriting) is # performing a lot of I/O against the disk, in some linux configurations
# Redis may block too long on the fsync() call. Note that there is no fix for # this currently, as even performing fsync in a different thread will block #
our synchronous write(2) call. # # In order to mitigate this problem it's possible to use the following option # that will prevent
fsync() from being called in the main process while a # BGSAVE or BGREWRITEAOF is in progress. # # This means that while another
child is saving the durability of Redis is # the same as "appendfsync none", that in pratical terms means that it is # possible to lost
up to 30 seconds of log in the worst scenario (with the # default Linux settings). # # If you have latency
problems turn this to "yes". Otherwise leave it as # "no" that is the safest pick from the point of view of durability. no-appendfsync-on-rewrite
no # Automatic rewrite of the append only file. # Redis is able to automatically rewrite the log file implicitly calling # BGREWRITEAOF when the
AOF log size will growth by the specified percentage. # # This is how it works: Redis remembers the size of the AOF file after the # latest rewrite (or if no
rewrite happened since the restart, the size of # the AOF at startup is used). # # This base size is compared to the current size. If the current
size is # bigger than the specified percentage, the rewrite is triggered. Also # you need to specify a minimal size for the AOF file to be
rewritten, this # is useful to avoid rewriting the AOF file even if the percentage increase # is reached but it is still pretty
small. # # Specify a precentage of zero in order to disable the automatic AOF # rewrite feature. auto-aof-rewrite-percentage 100 auto-aof-rewrite-min-size 64mb
################################## SLOW LOG ################################### # The Redis Slow Log is a system to log queries that exceeded a specified # execution time. The execution time does not include
the I/O operations # like talking with the client, sending the reply and so forth, # but just the time needed to actually execute the command (this is the
only # stage of command execution where the thread is blocked and can not serve # other requests in the meantime). # # You can configure the slow
log with two parameters: one tells Redis # what is the execution time, in microseconds, to exceed in order for the
# command to get logged, and the other parameter is the length of the # slow log. When a new command is logged
the oldest one is removed from the # queue of logged commands. # The following time is expressed in microseconds, so 1000000 is equivalent
# to one second. Note that a negative number disables the slow log, while # a value of zero forces the logging of every command. slowlog-log-slower-than 10000 #
There is no limit to this length. Just be aware that it will consume memory. # You can reclaim memory used by the slow log with SLOWLOG RESET. slowlog-max-len 1024 ################################
VIRTUAL MEMORY ############################### ### WARNING! Virtual Memory is deprecated in Redis 2.4 ### The use of Virtual Memory is strongly
discouraged. ### WARNING! Virtual Memory is deprecated in Redis 2.4 ### The use of Virtual Memory is strongly discouraged. #
Virtual Memory allows Redis to work with datasets bigger than the actual # amount of RAM needed to hold the whole dataset in memory. # In order to do
so very used keys are taken in memory while the other keys # are swapped into a swap file, similarly to what operating systems do # with memory pages.
# # To enable VM just set 'vm-enabled' to yes, and set the following three # VM parameters accordingly to your needs. vm-enabled no 指定是否启用虚拟内存机制，默认值为no，简单的介绍一下，VM机制将数据分页存放，由Redis将访问量较少的页即冷数据swap到磁盘上，访问多的页面由磁盘自动换出到内存中（在后面的文章我会仔细分析Redis的VM机制）
# vm-enabled yes # This is the path of the Redis swap file. As you can guess, swap files # can't be shared by different Redis instances, so make sure to use a
swap # file for every redis process you are running. Redis will complain if the # swap file is already in use.
# # The best kind of storage for the Redis swap file (that's accessed at random) # is a Solid State Disk (SSD). # # *** WARNING *** if you
are using a shared hosting the default of putting # the swap file under /tmp is not secure. Create a dir with access granted
# only to Redis user and configure Redis to create the swap file there. vm-swap-file /tmp/redis.swap 虚拟内存文件路径，默认值为/tmp/redis.swap，不可多个Redis实例共享 #
vm-max-memory configures the VM to use at max the specified amount of # RAM. Everything that deos not fit will be swapped on disk *if*
possible, that # is, if there is still enough contiguous space in the swap file. # # With vm-max-memory 0 the
system will swap everything it can. Not a good # default, just specify the max amount of RAM you can in bytes, but it's # better to leave
some margin. For instance specify an amount of RAM # that's more or less between 60 and 80% of your
free RAM. vm-max-memory 0 将所有大于vm-max-memory的数据存入虚拟内存,无论vm-max-memory设置多小,所有索引数据都是内存存储的(Redis的索引数据 就是keys),也就是说,当vm-max-memory设置为0的时候,其实是所有value都存在于磁盘。默认值为0 # Redis swap files is split
into pages. An object can be saved using multiple # contiguous pages, but pages can't be shared between different objects. # So if your page is too big, small objects swapped out on disk
will waste # a lot of space. If you page is too small, there is less space in the swap # file (assuming you configured the same
number of total swap file pages). # # If you use a lot of small objects, use a page size of 64 or 32 bytes.
# If you use a lot of big objects, use a bigger page size. # If unsure, use the default :) vm-page-size 32 Redis
swap文件分成了很多的page，一个对象可以保存在多个page上面，但一个page上不能被多个对象共享，vm-page-size是要根据存储的 数据大小来设定的，作者建议如果存储很多小对象，page大小最好设置为32或者64bytes；如果存储很大大对象，则可以使用更大的page，如果不 确定，就使用默认值 # Number of total memory pages in the
swap file. # Given that the page table (a bitmap of free/used pages) is taken in memory, # every 8 pages on disk will consume 1 byte of RAM.
# # The total swap size is vm-page-size * vm-pages # # With the default of 32-bytes memory pages and 134217728 pages
Redis will # use a 4 GB swap file, that will use 16 MB of RAM for the page table. # # It's better to use the
smallest acceptable value for your application, # but the default is large in order to work in most conditions. vm-pages 134217728 设置swap文件中的page数量，由于页表（一种表示页面空闲或使用的bitmap）是在放在内存中的，，在磁盘上每8个pages将消耗1byte的内存。
# Max number of VM I/O threads running at the same time. # This threads are used to read/write data from/to swap file, since they # also encode and decode
objects from disk to memory or the reverse, a bigger # number of threads can help with big objects even if they can't help with #
I/O itself as the physical device may not be able to couple with many # reads/writes operations at the same time. # # The special value of 0 turn
off threaded I/O and enables the blocking # Virtual Memory implementation. vm-max-threads 4 设置访问swap文件的线程数,最好不要超过机器的核数,如果设置为0,那么所有对swap文件的操作都是串行的，可能会造成比较长时间的延迟。默认值为4 ###############################
ADVANCED CONFIG ############################### # Hashes are encoded in a special way (much more memory efficient) when they # have at max a given numer of elements, and the
biggest element does not # exceed a given threshold. You can configure this limits with the following # configuration directives. hash-max-zipmap-entries 512 hash-max-zipmap-value 64 指定在超过一定的数量或者最大的元素超过某一临界值时，采用一种特殊的哈希算法
# Similarly to hashes, small lists are also encoded in a special way in order # to save a lot of space. The special representation is only
used when # you are under the following limits: list-max-ziplist-entries 512 list-max-ziplist-value 64 # Sets have a special encoding in just one case: when a
set is composed # of just strings that happens to be integers in radix 10 in the range # of 64 bit signed integers.
# The following configuration setting sets the limit in the size of the # set in order to use this special memory saving encoding.
set-max-intset-entries 512 # Similarly to hashes and lists, sorted sets are also specially encoded in # order to save a lot of space.
This encoding is only used when the length and # elements of a sorted set are below the following limits: zset-max-ziplist-entries 128 zset-max-ziplist-value 64 #
Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in # order to help rehashing the main Redis hash table
(the one mapping top-level # keys to values). The hash table implementation redis uses (see dict.c) # performs a lazy rehashing: the more operation you run into an hash table # that is rhashing, the more rehashing "steps" are
performed, so if the # server is idle the rehashing is never complete and some more memory is used # by the hash table. # # The default is to use this
millisecond 10 times every second in order to # active rehashing the main dictionaries, freeing memory when possible. # # If unsure: # use "activerehashing
no" if you have hard latency requirements and it is # not a good thing in your environment that Redis can reply form time to time # to queries with 2 milliseconds
delay. # # use "activerehashing yes" if you don't have such hard requirements but # want to free memory asap when possible.
activerehashing yes ################################## INCLUDES ################################### # Include one or more other config files here. This is useful if you # have a standard template
that goes to all redis server but also need # to customize a few per-server settings. Include files can include # other files, so use this wisely. # # include /path/to/local.conf
# include /path/to/other.conf

版权声明：本文为博主原创文章，未经博主允许不得转载。请多多指教~
关键词：