你的位置:首页 > 数据库

[数据库]Redis配置文件(redis.conf)说明


    # Redis configuration file example 
     
    # Note on units: when memory size is needed, it is possible to specify 
    # it in the usual form of 1k 5GB 4M and so forth: 
    # 内存大小的配置,下面是内存大小配置的转换方式 
    # 
    # 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. 
    # 内存大小的配置,不区分大小写 
     
    ################################## 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. 
    # 
    # Notice option "include" won't be rewritten by command "CONFIG REWRITE" 
    # from admin or Redis Sentinel. Since Redis always uses the last processed 
    # line as value of a configuration directive, you'd better put includes 
    # at the beginning of this file to avoid overwriting config change at runtime. 
    # 
    # If instead you are interested in using includes to override configuration 
    # options, it is better to use include as the last line. 
    # 
    # include /path/to/local.conf 
    # include /path/to/other.conf 
    # 当配置多个redis时,可能大部分配置一样,而对于不同的redis,只有少部分配置需要定制 
    # 就可以配置一个公共的模板配置。 
    # 对于具体的reids,只需设置少量的配置,并用include把模板配置包含进来即可。 
    # 
    # 值得注意的是,对于同一个配置项,redis只对最后一行的有效 
    # 所以为避免模板配置覆盖当前配置,应在配置文件第一行使用include 
    # 当然,如果模板配置的优先级比较高,就在配置文件最后一行使用include 
     
    ################################ GENERAL  ##################################### 
     
    # 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. 
    # yes为使用守护进程,此时redis的进程ID会被写进 pidfile的配置中 
    daemonize yes 
     
    # When running daemonized, Redis writes a pid file in /var/run/redis.pid by 
    # default. You can specify a custom pid file location here. 
    # 当redis以守护进程的方式启动时,redis的进程ID将会写在这个文件中
                # 该文件目录相对于配置文件目录,当文件不存在时,不会主动创建,需要手动创建该文件
    pidfile /var/run/redis.pid 
     
    # Accept connections on the specified port, default is 6379. 
    # If port 0 is specified Redis will not listen on a TCP socket. 
    # redis 启动的端口。【应该知道redis是服务端吧】 
    port 6379 
     
    # TCP listen() backlog. 
    # 
    # In high requests-per-second environments you need an high backlog in order 
    # to avoid slow clients connections issues. Note that the Linux kernel 
    # will silently truncate it to the value of /proc/sys/net/core/somaxconn so 
    # make sure to raise both the value of somaxconn and tcp_max_syn_backlog 
    # in order to get the desired effect. 
    # 最大链接缓冲池的大小,这里应该是指的未完成链接请求的数量 
    #(测试值为1时,仍可以有多个链接) 
    # 但该值与listen函数中的backlog意义应该是相同的,源码中该值就是被用在了listen函数中 
    # 该值同时受/proc/sys/net/core/somaxconn 和 tcp_max_syn_backlog(/etc/sysctl.conf中配置)的限制 
    # tcp_max_syn_backlog 指的是未完成链接的数量 
    tcp-backlog 511 
     
    # By default Redis listens for connections from all the network interfaces 
    # available on the server. It is possible to listen to just one or multiple 
    # interfaces using the "bind" configuration directive, followed by one or 
    # more IP addresses. 
    # 绑定ip,指定ip可以连接到redis 
    # 
    # Examples: 
    # 
    # bind 192.168.1.100 10.0.0.1 
    # bind 127.0.0.1 
     
    # 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. 
    # 
    # 这个应该就是以文件形式创建的socket 
    # unixsocket /tmp/redis.sock 
    # unixsocketperm 755 
     
    # Close the connection after a client is idle for N seconds (0 to disable) 
    # 超时断链机制,如果一个链接在N秒内没有任何操作,则断开该链接 
    # N为0时,该机制失效 
    timeout 0 
     
    # TCP keepalive. 
    # 
    # If non-zero, use SO_KEEPALIVE to send TCP ACKs to clients in absence 
    # of communication. This is useful for two reasons: 
    # 
    # 1) Detect dead peers. 
    # 2) Take the connection alive from the point of view of network 
    #    equipment in the middle. 
    # 
    # On Linux, the specified value (in seconds) is the period used to send ACKs. 
    # Note that to close the connection the double of the time is needed. 
    # On other kernels the period depends on the kernel configuration. 
    # 就像心跳检测一样,检查链接是否保持正常,同时也可以保持正常链接的通信 
    # 建议值为60 
    # 
    # A reasonable value for this option is 60 seconds. 
    tcp-keepalive 0 
     
    # Specify the server verbosity level. 
    # This 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) 
    # 日志级别 
    loglevel notice 
     
    # Specify the log file name. Also the empty string 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 
    # 日志存放路径,默认是输出到标准输出,但当以守护进程方式启动时,默认输出到/dev/null(传说中的linux黑洞) 
    logfile "" 
     
    # To enable logging to the system logger, just set 'syslog-enabled' to yes, 
    # and optionally update the other syslog parameters to suit your needs. 
    # yes 表示将日志写到系统日志中 
    # syslog-enabled no 
     
    # Specify the syslog identity. 
    # 当syslog-enabled为yes时,指定系统日志的标示为 redis 
    # syslog-ident redis 
     
    # Specify the syslog facility. Must be USER or between LOCAL0-LOCAL7. 
    # 指定系统日志的设备 
    # 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 
    # redis的数据库格式,默认16个(0~15),默认使用第0个。 
    databases 16 
     
    ################################ 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. 
    # 快照,即将数据写到硬盘上,在<seconds>秒内,至少有<changes>次写入数据库操作 
    # 则会将数据写入硬盘一次。 
    # 将save行注释掉则永远不会写入硬盘 
    # save "" 表示删除所有的快照点 
    # 
    #   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. 
    # 
    #   It is also possible to remove all the previously configured save 
    #   points by adding a save directive with a single empty string argument 
    #   like in the following example: 
    # 
    #   save "" 
     
    save 900 1 
    save 300 10 
    save 60 10000 
     
    # By default Redis will stop accepting writes if RDB snapshots are enabled 
    # (at least one save point) and the latest background save failed. 
    # This will make the user aware (in a hard way) that data is not persisting 
    # on disk properly, otherwise chances are that no one will notice and some 
    # disaster will happen. 
    # 
    # If the background saving process will start working again Redis will 
    # automatically allow writes again. 
    # 
    # However if you have setup your proper monitoring of the Redis server 
    # and persistence, you may want to disable this feature so that Redis will 
    # continue to work as usual even if there are problems with disk, 
    # permissions, and so forth. 
    # 当做快照失败的时候,redis会停止继续向其写入数据,保证第一时间发现redis快照出现问题 
    # 当然,通过下面配置为 no,即使redis快照失败,也能继续向redis写入数据 
    stop-writes-on-bgsave-error yes 
     
    # 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. 
    # 快照的时候,是否用LZF压缩,使用压缩会占一定的cpu,但不使用压缩,快照会很大 
    rdbcompression yes 
     
    # Since version 5 of RDB a CRC64 checksum is placed at the end of the file. 
    # This makes the format more resistant to corruption but there is a performance 
    # hit to pay (around 10%) when saving and loading RDB files, so you can disable it 
    # for maximum performances. 
    # 
    # RDB files created with checksum disabled have a checksum of zero that will 
    # tell the loading code to skip the check. 
    # 数据校验,快照末尾会存放一个校验值,保证数据的准确性 
    # 但数据校验会使性能下降约10%,默认开启校验 
    rdbchecksum yes 
     
    # The filename where to dump the DB 
    # 快照的名字 
    dbfilename dump.rdb 
     
    # The working directory. 
    # 
    # The DB will be written inside this directory, with the filename specified 
    # above using the 'dbfilename' configuration directive. 
    #  
    # The Append Only File will also be created inside this directory. 
    #  
    # Note that you must specify a directory here, not a file name. 
    #  
    # 快照存放的目录 
    # linux root下测试,会发现该进程会在当前目录下创建一个dump.rdb 
    # 但快照却放在了根目录/下,重启的时候,是不会从快照中恢复数据的 
    # 当把根目录下的dump.rdb文件拷贝到当前目录的时候,再次启动,就会从快照中恢复数据 
    # 而且以后的快照也都在当前目录的dump.rdb中做操作 
    # 
    # 值得一提的是,快照是异步方式的,如果在还未达到快照的时候,修改了数据,而且redis发生问题crash了 
    # 那么中间的修改数据是不会被保存到dump.rdb快照中的 
    # 解决办法就是用Append Only Mode的同步模式(下面将会有该配置项) 
    # 将会把每个操作写到Append Only File中,该文件也存放于当前配置的目录 
    # 建议使用绝对路径!!! 
    #  
    dir ./ 
     
    ################################# 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. 
    # 
    # 主从复制,类似于双机备份。 
    # 配置需指定主机的ip 和port 
    # slaveof <masterip> <masterport> 
     
    # 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. 
    # 
    # 如果主机redis需要密码,则指定密码 
    # 密码配置在下面安全配置中 
    # masterauth <master-password> 
     
    # When a slave loses its 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 date 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. 
    # 
    # 当从机与主机断开时,即同步出现问题的时候,从机有两种处理方式 
    # yes, 继续响应客户端请求,但可能有脏数据(过期数据、空数据等) 
    # no,对客户端的请求统一回复为“SYNC with master in progress”,除了INFO和SLAVEOF命令 
    slave-serve-stale-data yes 
     
    # You can configure a slave instance to accept writes or not. Writing against 
    # a slave instance may be useful to store some ephemeral data (because data 
    # written on a slave will be easily deleted after resync with the master) but 
    # may also cause problems if clients are writing to it because of a 
    # misconfiguration. 
    # 
    # Since Redis 2.6 by default slaves are read-only. 
    # 
    # Note: read only slaves are not designed to be exposed to untrusted clients 
    # on the internet. It's just a protection layer against misuse of the instance. 
    # Still a read only slave exports by default all the administrative commands 
    # such as CONFIG, DEBUG, and so forth. To a limited extent you can improve 
    # security of read only slaves using 'rename-command' to shadow all the 
    # administrative / dangerous commands. 
    # slave只读选项,设置从机只读(默认)。 
    # 即使设置可写,当下一次从主机上同步数据,仍然会删除当前从机上写入的数据 
    # 【待测试】:主机与从机互为slave会出现什么情况? 
    # 【预期三种结果】:1. 提示报错 2. 主从服务器数据不可控 3. 一切正常 
    slave-read-only 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. 
    # 
    # 从服务器向主服务器发送心跳包,默认10发送一次 
    # repl-ping-slave-period 10 
     
    # The following option sets the replication timeout for: 
    # 
    # 1) Bulk transfer I/O during SYNC, from the point of view of slave. 
    # 2) Master timeout from the point of view of slaves (data, pings). 
    # 3) Slave timeout from the point of view of masters (REPLCONF ACK pings). 
    # 
    # 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. 
    # 
    # 超时响应时间,值必须比repl-ping-slave-period大 
    # 批量数据传输超时、ping超时 
    # repl-timeout 60 
     
    # Disable TCP_NODELAY on the slave socket after SYNC? 
    # 
    # If you select "yes" Redis will use a smaller number of TCP packets and 
    # less bandwidth to send data to slaves. But this can add a delay for 
    # the data to appear on the slave side, up to 40 milliseconds with 
    # Linux kernels using a default configuration. 
    # 
    # If you select "no" the delay for data to appear on the slave side will 
    # be reduced but more bandwidth will be used for replication. 
    # 
    # By default we optimize for low latency, but in very high traffic conditions 
    # or when the master and slaves are many hops away, turning this to "yes" may 
    # be a good idea. 
    # 主从同步是否延迟 
    # yes 有延迟,约40毫秒(linux kernel的默认配置),使用较少的数据包,较小的带宽 
    # no 无延迟(减少延迟),但需要更大的带宽 
    repl-disable-tcp-nodelay no 
     
    # Set the replication backlog size. The backlog is a buffer that accumulates 
    # slave data when slaves are disconnected for some time, so that when a slave 
    # wants to reconnect again, often a full resync is not needed, but a partial 
    # resync is enough, just passing the portion of data the slave missed while 
    # disconnected. 
    # 
    # The biggest the replication backlog, the longer the time the slave can be 
    # disconnected and later be able to perform a partial resynchronization. 
    # 
    # The backlog is only allocated once there is at least a slave connected. 
    # 
    # 默认情况下,当slave重连的时候,会进行全量数据同步 
    # 但实际上slave只需要部分同步即可,这个选项设置部分同步的大小 
    # 设置值越大,同步的时间就越长 
    # repl-backlog-size 1mb 
     
    # After a master has no longer connected slaves for some time, the backlog 
    # will be freed. The following option configures the amount of seconds that 
    # need to elapse, starting from the time the last slave disconnected, for 
    # the backlog buffer to be freed. 
    # 
    # A value of 0 means to never release the backlog. 
    # 
    # 主机的后台日志释放时间,即当没有slave连接时,过多久释放后台日志 
    # 0表示不释放 
    # repl-backlog-ttl 3600 
     
    # The slave priority is an integer number published by Redis in the INFO output. 
    # It is used by Redis Sentinel in order to select a slave to promote into a 
    # master if the master is no longer working correctly. 
    # 
    # A slave with a low priority number is considered better for promotion, so 
    # for instance if there are three slaves with priority 10, 100, 25 Sentinel will 
    # pick the one with priority 10, that is the lowest. 
    # 
    # However a special priority of 0 marks the slave as not able to perform the 
    # role of master, so a slave with priority of 0 will never be selected by 
    # Redis Sentinel for promotion. 
    # 
    # By default the priority is 100. 
    # 当主机crash的时候,在从机中选择一台作为主机,数字越小,优先级越高 
    # 0 表示永远不作为主机,默认值是100 
    slave-priority 100 
     
    # It is possible for a master to stop accepting writes if there are less than 
    # N slaves connected, having a lag less or equal than M seconds. 
    # 
    # The N slaves need to be in "online" state. 
    # 
    # The lag in seconds, that must be <= the specified value, is calculated from 
    # the last ping received from the slave, that is usually sent every second. 
    # 
    # This option does not GUARANTEES that N replicas will accept the write, but 
    # will limit the window of exposure for lost writes in case not enough slaves 
    # are available, to the specified number of seconds. 
    # 
    # For example to require at least 3 slaves with a lag <= 10 seconds use: 
    # 
    # 当slave数量小于min-slaves-to-write,且延迟小于等于min-slaves-max-lag时, 
    # 主机停止写入操作 
    # 0表示禁用 
    # 默认min-slaves-to-write为0,即禁用。min-slaves-max-lag为10 
    # min-slaves-to-write 3 
    # min-slaves-max-lag 10 
    # 
    # Setting one or the other to 0 disables the feature. 
    # 
    # By default min-slaves-to-write is set to 0 (feature disabled) and 
    # min-slaves-max-lag is set to 10. 
     
    ################################## 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. 
    # 
    # redis密码,默认不配置,即无密码 
    # 这里注意,如果设置了密码,应该设置一个复杂度比较高的密码 
    # 因为redis的速度很快,每秒可以尝试150k次的密码测试,很容易对其进行暴力破解(跑码)。 
    # 疑问:这里为什么不设置一个针对主机的测试次数限制的,例如每10次,则禁止建立连接1个小时! 
    # requirepass foobared 
     
    # Command renaming. 
    # 
    # It is possible to change the name of dangerous commands in a shared 
    # environment. For instance the CONFIG command may be renamed into something 
    # hard to guess so that it will still be available for internal-use tools 
    # but not available for general clients. 
    # 
    # 命令重命名,将命令重命名为另一个字符串标识 
    # 如果命令为空串(""),则会彻底禁用该命令 
    # 命令重命名,会对写AOF(Append of file)文件、slave从机造成一些问题 
    # Example: 
    # 
    # rename-command CONFIG b840fc02d524045429941cc15f59e41cb7be6c52 
    # 
    # It is also possible to completely kill a command by renaming it into 
    # an empty string: 
    # 
    # rename-command CONFIG "" 
    # 
    # Please note that changing the name of commands that are logged into the 
    # AOF file or transmitted to slaves may cause problems. 
     
    ################################### LIMITS #################################### 
     
    # Set the max number of connected clients at the same time. By default 
    # this limit is set to 10000 clients, however if the Redis server is not 
    # able to configure the process file limit to allow for the specified limit 
    # the max number of allowed clients is set to the current file limit 
    # minus 32 (as Redis reserves a few file descriptors for internal uses). 
    # 
    # Once the limit is reached Redis will close all the new connections sending 
    # an error 'max number of clients reached'. 
    # 
    # 这只redis的最大连接数目,默认设置为10000个客户端 
    # 当超过限制时,将段开新的连接,并响应“max number of clients reached” 
    # maxclients 10000 
     
    # Don't use more memory than the specified amount of bytes. 
    # When the memory limit is reached Redis will try to remove keys 
    # according to the eviction policy selected (see maxmemory-policy). 
    # 
    # If Redis can't remove keys according to the policy, or if the policy is 
    # set to 'noeviction', Redis will start to reply with errors to commands 
    # that would use more memory, like SET, LPUSH, and so on, and will continue 
    # to reply to read-only commands like GET. 
    # 
    # This option is usually useful when using Redis as an LRU cache, or to set 
    # a hard memory limit for an instance (using the 'noeviction' policy). 
    # 
    # WARNING: If you have slaves attached to an instance with maxmemory on, 
    # the size of the output buffers needed to feed the slaves are subtracted 
    # from the used memory count, so that network problems / resyncs will 
    # not trigger a loop where keys are evicted, and in turn the output 
    # buffer of slaves is full with DELs of keys evicted triggering the deletion 
    # of more keys, and so forth until the database is completely emptied. 
    # 
    # In short... if you have slaves attached it is suggested that you set a lower 
    # limit for maxmemory so that there is some free RAM on the system for slave 
    # output buffers (but this is not needed if the policy is 'noeviction'). 
    # 
    # redis的最大内存限制,如果达到最大内存,会按照下面的maxmemory-policy进行清除 
    # 如果不能再清除或者maxmemory-policy为noeviction,则对于需要增加空间的操作,将会返回错误 
    # maxmemory <bytes> 
     
    # MAXMEMORY POLICY: how Redis will select what to remove when maxmemory 
    # is reached. You can select among five behaviors: 
    #  
    # 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 any of the above 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: 
    # 
    # 内存删除策略,默认volatile-lru,利用LRU算法,删除过期的key 
    # 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. 
    # 
    # LRU算法与最小TTL算法只是相对精确的算法,并不是绝对精确的算法 
    # 为了更精确,可以设置样本个数 
    # 比如设置3个样本,redis会选取三个key,并选择删除那个上次使用时间最远的 
    # maxmemory-samples 3 
     
    ############################## APPEND ONLY MODE ############################### 
     
    # By default Redis asynchronously dumps the dataset on disk. This mode is 
    # good enough in many applications, but an issue with the Redis process or 
    # a power outage may result into a few minutes of writes lost (depending on 
    # the configured save points). 
    # 
    # The Append Only File is an alternative persistence mode that provides 
    # much better durability. For instance using the default data fsync policy 
    # (see later in the config file) Redis can lose just one second of writes in a 
    # dramatic event like a server power outage, or a single write if something 
    # wrong with the Redis process itself happens, but the operating system is 
    # still running correctly. 
    # 
    # AOF and RDB persistence can be enabled at the same time without problems. 
    # If the AOF is enabled on startup Redis will load the AOF, that is the file 
    # with the better durability guarantees. 
    # 
    # Please check http://redis.io/topics/persistence for more information. 
    # 将对redis所有的操作都保存到AOF文件中 
    # 因为dump.rdb是异步的,在下次快照到达之前,如果出现crash等问题,会造成数据丢失 
    # 而AOF文件时同步记录的,所以会完整的恢复数据 
     
    appendonly no 
     
    # The name of the append only file (default: "appendonly.aof") 
    # AOF文件的名字 
     
    appendfilename "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 one time every second. Compromise. 
    # 
    # The default is "everysec", as 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 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. 
    # 
    # More details please check the following article: 
    # http://antirez.com/post/redis-persistence-demystified.html 
    # 
    # If unsure, use "everysec". 
    # redis的数据同步方式,三种 
    # no,redis本身不做同步,由OS来做。redis的速度会很快 
    # always,在每次写操作之后,redis都进行同步,即写入AOF文件。redis会变慢,但是数据更安全 
    # everysec,折衷考虑,每秒同步一次数据。【默认】 
     
    # appendfsync always 
    appendfsync everysec 
    # appendfsync no 
     
    # 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". In practical terms, this means that it is 
    # possible to lose 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. 
    # redis的同步方式中,always和everysec,快照和写AOF可能会执行大量的硬盘I/O操作, 
    # 而在一些Linux的配置中,redis会阻塞很久,而redis本身并没有很好的解决这一问题。 
    # 为了缓和这一问题,redis提供no-appendfsync-on-rewrite选项, 
    # 即当有另外一个进程在执行保存操作的时候,redis采用no的同步方式。 
    # 最坏情况下会有延迟30秒的同步延迟。 
    # 如果你觉得这样做会有潜在危险,则请将该选项改为yes。否则就保持默认值no(基于稳定性考虑)。 
     
    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 grows by the specified percentage. 
    #  
    # This is how it works: Redis remembers the size of the AOF file after the 
    # latest rewrite (if no rewrite has 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 percentage of zero in order to disable the automatic AOF 
    # rewrite feature. 
    # 自动重写AOF文件 
    # 当AOF日志文件大小增长到指定百分比时,redis会自动隐式调用BGREWRITEAOF来重写AOF文件 
    # redis会记录上次重写AOF文件之后的大小, 
    # 如果当前文件大小增加了auto-aof-rewrite-percentage,则会触发重写AOF日志功能 
    # 当然如果文件过小,比如小于auto-aof-rewrite-min-size这个大小,是不会触发重写AOF日志功能的 
    # auto-aof-rewrite-percentage为0时,禁用重写功能 
     
    auto-aof-rewrite-percentage 100 
    auto-aof-rewrite-min-size 64mb 
     
    ################################ LUA SCRIPTING  ############################### 
     
    # Max execution time of a Lua script in milliseconds. 
    # 
    # If the maximum execution time is reached Redis will log that a script is 
    # still in execution after the maximum allowed time and will start to 
    # reply to queries with an error. 
    # 
    # When a long running script exceed the maximum execution time only the 
    # SCRIPT KILL and SHUTDOWN NOSAVE commands are available. The first can be 
    # used to stop a script that did not yet called write commands. The second 
    # is the only way to shut down the server in the case a write commands was 
    # already issue by the script but the user don't want to wait for the natural 
    # termination of the script. 
    # 
    # Set it to 0 or a negative value for unlimited execution without warnings. 
    # LUA脚本的最大执行时间(单位是毫秒),默认5000毫秒,即5秒 
    # 如果LUA脚本执行超过这个限制,可以调用SCRIPT KILL和SHUTDOWN NOSAVE命令。 
    # SCRIPT KILL可以终止脚本执行 
    # SHUTDOWN NOSAVE关闭服务,防止LUA脚本的写操作发生 
    # 该值为0或者负数,表示没有限制时间 
    lua-time-limit 5000 
     
    ################################## 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. 
    # 记录执行比较慢的命令 
    # 执行比较慢仅仅是指命令的执行时间,不包括客户端的链接与响应等时间 
    # slowlog-log-slower-than 设定这个慢的时间,单位是微妙,即1000000表示1秒,0表示所有命令都记录,负数表示不记录 
    # slowlog-max-len表示记录的慢命令的个数,超过限制,则最早记录的命令会被移除 
    # 命令的长度没有限制,但是会消耗内存,用SLOWLOG RESET来收回这些消耗的内存 
     
    # 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 128 
     
    ################################ LATENCY MONITOR ############################## 
     
    # The Redis latency monitoring subsystem samples different operations 
    # at runtime in order to collect data related to possible sources of 
    # latency of a Redis instance. 
    # 
    # Via the LATENCY command this information is available to the user that can 
    # print graphs and obtain reports. 
    # 
    # The system only logs operations that were performed in a time equal or 
    # greater than the amount of milliseconds specified via the 
    # latency-monitor-threshold configuration directive. When its value is set 
    # to zero, the latency monitor is turned off. 
    # 
    # By default latency monitoring is disabled since it is mostly not needed 
    # if you don't have latency issues, and collecting data has a performance 
    # impact, that while very small, can be measured under big load. Latency 
    # monitoring can easily be enalbed at runtime using the command 
    # "CONFIG SET latency-monitor-threshold <milliseconds>" if needed. 
    # 延迟监控器 
    # redis延迟监控子系统在运行时,会抽样检测可能导致延迟的不同操作 
    # 通过LATENCY命令可以打印相关信息和报告, 命令如下(摘自源文件注释): 
    # LATENCY SAMPLES: return time-latency samples for the specified event. 
    # LATENCY LATEST: return the latest latency for all the events classes. 
    # LATENCY DOCTOR: returns an human readable analysis of instance latency. 
    # LATENCY GRAPH: provide an ASCII graph of the latency of the specified event. 
    #  
    # 系统只记录超过设定值的操作,单位是毫秒,0表示禁用该功能 
    # 可以通过命令“CONFIG SET latency-monitor-threshold <milliseconds>” 直接设置而不需要重启redis 
     
    latency-monitor-threshold 0 
     
    ############################# Event notification ############################## 
     
    # Redis can notify Pub/Sub clients about events happening in the key space. 
    # This feature is documented at http://redis.io/topics/keyspace-events 
    #  
    # For instance if keyspace events notification is enabled, and a client 
    # performs a DEL operation on key "foo" stored in the Database 0, two 
    # messages will be published via Pub/Sub: 
    # 
    # PUBLISH __keyspace@0__:foo del 
    # PUBLISH __keyevent@0__:del foo 
    # 
    # It is possible to select the events that Redis will notify among a set 
    # of classes. Every class is identified by a single character: 
    # 
    #  K     Keyspace events, published with __keyspace@<db>__ prefix. 
    #  E     Keyevent events, published with __keyevent@<db>__ prefix. 
    #  g     Generic commands (non-type specific) like DEL, EXPIRE, RENAME, ... 
    #  $     String commands 
    #  l     List commands 
    #  s     Set commands 
    #  h     Hash commands 
    #  z     Sorted set commands 
    #  x     Expired events (events generated every time a key expires) 
    #  e     Evicted events (events generated when a key is evicted for maxmemory) 
    #  A     Alias for g$lshzxe, so that the "AKE" string means all the events. 
    # 
    #  The "notify-keyspace-events" takes as argument a string that is composed 
    #  by zero or multiple characters. The empty string means that notifications 
    #  are disabled at all. 
    # 
    #  Example: to enable list and generic events, from the point of view of the 
    #           event name, use: 
    # 
    #  notify-keyspace-events Elg 
    # 
    #  Example 2: to get the stream of the expired keys subscribing to channel 
    #             name __keyevent@0__:expired use: 
    # 
    #  notify-keyspace-events Ex 
    # 
    #  By default all notifications are disabled because most users don't need 
    #  this feature and the feature has some overhead. Note that if you don't 
    #  specify at least one of K or E, no events will be delivered. 
    # 事件通知,当事件发生时,redis可以通知Pub/Sub客户端 
    # 空串表示禁用事件通知 
    # 注意:K和E至少要指定一个,否则不会有事件通知 
    notify-keyspace-events "" 
     
    ############################### ADVANCED CONFIG ############################### 
     
    # Hashes are encoded using a memory efficient data structure when they have a 
    # small number of entries, and the biggest entry does not exceed a given 
    # threshold. These thresholds can be configured using the following directives. 
    # 当hash数目比较少,并且最大元素没有超过给定值时,Hash使用比较有效的内存数据结构来存储。 
    # 即ziplist的结构(压缩的双向链表),参考:http://blog.csdn.net/benbendy1984/article/details/7796956 
    hash-max-ziplist-entries 512 
    hash-max-ziplist-value 64 
     
    # 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配置同Hash 
    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. 
    # Sets的元素如果全部是整数(10进制),且为64位有符号整数,则采用特殊的编码方式。 
    # 其元素个数限制配置如下: 
    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: 
    # sorted set 同Hash和List 
    zset-max-ziplist-entries 128 
    zset-max-ziplist-value 64 
     
    # HyperLogLog sparse representation bytes limit. The limit includes the 
    # 16 bytes header. When an HyperLogLog using the sparse representation crosses 
    # this limit, it is converted into the dense representation. 
    # 
    # A value greater than 16000 is totally useless, since at that point the 
    # dense representation is more memory efficient. 
    #  
    # The suggested value is ~ 3000 in order to have the benefits of 
    # the space efficient encoding without slowing down too much PFADD, 
    # which is O(N) with the sparse encoding. The value can be raised to 
    # ~ 10000 when CPU is not a concern, but space is, and the data set is 
    # composed of many HyperLogLogs with cardinality in the 0 - 15000 range. 
    # 关于HyperLogLog的介绍:http://www.redis.io/topics/data-types-intro#hyperloglogs 
    # HyperLogLog稀疏表示限制设置,如果其值大于16000,则仍然采用稠密表示,因为这时稠密表示更能有效使用内存 
    # 建议值为3000 
    hll-sparse-max-bytes 3000 
     
    # 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 a hash table 
    # that is rehashing, 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. 
    # 每100毫秒,redis将用1毫秒的时间对Hash表进行重新Hash。 
    # 采用懒惰Hash方式:操作Hash越多,则重新Hash的可能越多,若根本就不操作Hash,则不会重新Hash 
    # 默认每秒10次重新hash主字典,释放可能释放的内存 
    # 重新hash会造成延迟,如果对延迟要求较高,则设为no,禁止重新hash。但可能会浪费很多内存 
    activerehashing yes 
     
    # The client output buffer limits can be used to force disconnection of clients 
    # that are not reading data from the server fast enough for some reason (a 
    # common reason is that a Pub/Sub client can't consume messages as fast as the 
    # publisher can produce them). 
    # 
    # The limit can be set differently for the three different classes of clients: 
    # 
    # normal -> normal clients including MONITOR clients 
    # slave  -> slave clients 
    # pubsub -> clients subscribed to at least one pubsub channel or pattern 
    # 
    # The syntax of every client-output-buffer-limit directive is the following: 
    # 
    # 客户端输出缓冲区限制,当客户端从服务端的读取速度不够快时,则强制断开 
    # 三种不同的客户端类型:normal、salve、pubsub,语法如下: 
    # client-output-buffer-limit <class> <hard limit> <soft limit> <soft seconds> 
    # 
    # A client is immediately disconnected once the hard limit is reached, or if 
    # the soft limit is reached and remains reached for the specified number of 
    # seconds (continuously). 
    # So for instance if the hard limit is 32 megabytes and the soft limit is 
    # 16 megabytes / 10 seconds, the client will get disconnected immediately 
    # if the size of the output buffers reach 32 megabytes, but will also get 
    # disconnected if the client reaches 16 megabytes and continuously overcomes 
    # the limit for 10 seconds. 
    # 
    # By default normal clients are not limited because they don't receive data 
    # without asking (in a push way), but just after a request, so only 
    # asynchronous clients may create a scenario where data is requested faster 
    # than it can read. 
    # 
    # Instead there is a default limit for pubsub and slave clients, since 
    # subscribers and slaves receive data in a push fashion. 
    # 
    # Both the hard or the soft limit can be disabled by setting them to zero. 
    # 当达到硬限制,或者达到软限制且持续了算限制秒数,则立即与客户端断开 
    # 限制设为0表示禁止该功能 
    # 普通用户默认不限制 
    client-output-buffer-limit normal 0 0 0 
    client-output-buffer-limit slave 256mb 64mb 60 
    client-output-buffer-limit pubsub 32mb 8mb 60 
     
    # Redis calls an internal function to perform many background tasks, like 
    # closing connections of clients in timeout, purging expired keys that are 
    # never requested, and so forth. 
    # 
    # Not all tasks are performed with the same frequency, but Redis checks for 
    # tasks to perform accordingly to the specified "hz" value. 
    # 
    # By default "hz" is set to 10. Raising the value will use more CPU when 
    # Redis is idle, but at the same time will make Redis more responsive when 
    # there are many keys expiring at the same time, and timeouts may be 
    # handled with more precision. 
    # 
    # The range is between 1 and 500, however a value over 100 is usually not 
    # a good idea. Most users should use the default of 10 and raise this up to 
    # 100 only in environments where very low latency is required. 
    # redis调用内部函数执行的后台任务的频率 
    # 后台任务比如:清除过期数据、客户端超时链接等 
    # 默认为10,取值范围1~500, 
    # 对延迟要求很低的可以设置超过100以上 
    hz 10 
     
    # When a child rewrites the AOF file, if the following option is enabled 
    # the file will be fsync-ed every 32 MB of data generated. This is useful 
    # in order to commit the file to the disk more incrementally and avoid 
    # big latency spikes. 
    # 当修改AOF文件时,该设置为yes,则每生成32MB的数据,就进行同步 
    aof-rewrite-incremental-fsync yes