Redis配置文件詳解

# redis 配置文件示例  

   

# 當你須要爲某個配置項指定內存大小的時候，必需要帶上單位，  

# 一般的格式就是 1k 5gb 4m 等醬紫：  

#  

# 1k  => 1000 bytes  

# 1kb => 1024 bytes  

# 1m  => 1000000 bytes  

# 1mb => 1024*1024 bytes  

# 1g  => 1000000000 bytes  

# 1gb => 1024*1024*1024 bytes  

#  

# 單位是不區分大小寫的，你寫 1K 5GB 4M 也行  

   

################################## INCLUDES ###################################  

   

# 假如說你有一個可用於全部的 redis server 的標準配置模板，  

# 但針對某些 server 又須要一些個性化的設置，  

# 你可使用 include 來包含一些其餘的配置文件，這對你來講是很是有用的。  

#  

# 可是要注意哦，include 是不能被 config rewrite 命令改寫的  

# 因爲 redis 老是以最後的加工線做爲一個配置指令值，因此你最好是把 include 放在這個文件的最前面，  

# 以免在運行時覆蓋配置的改變，相反，你就把它放在後面（外國人真囉嗦）。  

#  

# include /path/to/local.conf  

# include /path/to/other.conf  

   

################################ 經常使用 #####################################  

   

# 默認狀況下 redis 不是做爲守護進程運行的，若是你想讓它在後臺運行，你就把它改爲 yes。  

# 當redis做爲守護進程運行的時候，它會寫一個 pid 到 /var/run/redis.pid 文件裏面。  

daemonize no  

   

# 當redis做爲守護進程運行的時候，它會把 pid 默認寫到 /var/run/redis.pid 文件裏面，  

# 可是你能夠在這裏本身制定它的文件位置。  

pidfile /var/run/redis.pid  

   

# 監聽端口號，默認爲 6379，若是你設爲 0 ，redis 將不在 socket 上監放任何客戶端鏈接。  

port 6379  

   

# TCP 監聽的最大容納數量  

#  

# 在高併發的環境下，你須要把這個值調高以免客戶端鏈接緩慢的問題。  

# Linux 內核會一言不發的把這個值縮小成 /proc/sys/net/core/somaxconn 對應的值，  

# 因此你要修改這兩個值才能達到你的預期。  

tcp-backlog 511  

   

# 默認狀況下，redis 在 server 上全部有效的網絡接口上監聽客戶端鏈接。  

# 你若是隻想讓它在一個網絡接口上監聽，那你就綁定一個IP或者多個IP。  

#  

# 示例，多個IP用空格隔開:  

#  

# bind 192.168.1.100 10.0.0.1  

# bind 127.0.0.1  

   

# 指定 unix socket 的路徑。  

#  

# unixsocket /tmp/redis.sock  

# unixsocketperm 755  

   

# 指定在一個 client 空閒多少秒以後關閉鏈接（0 就是無論它）  

timeout 0  

   

# tcp 心跳包。  

#  

# 若是設置爲非零，則在與客戶端缺少通信的時候使用 SO_KEEPALIVE 發送 tcp acks 給客戶端。  

# 這個之全部有用，主要由兩個緣由：  

#  

# 1) 防止死的 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.  

#  

# A reasonable value for this option is 60 seconds.  

# 推薦一個合理的值就是60秒  

tcp-keepalive 0  

   

# 定義日誌級別。  

# 能夠是下面的這些值：  

# debug (適用於開發或測試階段)  

# verbose (many rarely useful info, but not a mess like the debug level)  

# notice (適用於生產環境)  

# warning (僅僅一些重要的消息被記錄)  

loglevel notice  

   

# 指定日誌文件的位置  

logfile ""  

   

# 要想把日誌記錄到系統日誌，就把它改爲 yes，  

# 也能夠可選擇性的更新其餘的syslog 參數以達到你的要求  

# syslog-enabled no  

   

# 設置 syslog 的 identity。  

# syslog-ident redis  

   

# 設置 syslog 的 facility，必須是 USER 或者是 LOCAL0-LOCAL7 之間的值。  

# syslog-facility local0  

   

# 設置數據庫的數目。  

# 默認數據庫是 DB 0，你能夠在每一個鏈接上使用 select <dbid> 命令選擇一個不一樣的數據庫，  

# 可是 dbid 必須是一個介於 0 到 databasees - 1 之間的值  

databases 16  

   

################################ 快照 ################################  

#  

# 存 DB 到磁盤：  

#  

#   格式：save <間隔時間（秒）> <寫入次數>  

#  

#   根據給定的時間間隔和寫入次數將數據保存到磁盤  

#  

#   下面的例子的意思是：  

#   900 秒內若是至少有 1 個 key 的值變化，則保存  

#   300 秒內若是至少有 10 個 key 的值變化，則保存  

#   60 秒內若是至少有 10000 個 key 的值變化，則保存  

#　　  

#   注意：你能夠註釋掉全部的 save 行來停用保存功能。  

#   也能夠直接一個空字符串來實現停用：  

#   save ""  

   

save 900 1  

save 300 10  

save 60 10000  

   

# 默認狀況下，若是 redis 最後一次的後臺保存失敗，redis 將中止接受寫操做，  

# 這樣以一種強硬的方式讓用戶知道數據不能正確的持久化到磁盤，  

# 不然就會沒人注意到災難的發生。  

#  

# 若是後臺保存進程從新啓動工做了，redis 也將自動的容許寫操做。  

#  

# 然而你要是安裝了靠譜的監控，你可能不但願 redis 這樣作，那你就改爲 no 好了。  

stop-writes-on-bgsave-error yes  

   

# 是否在 dump .rdb 數據庫的時候使用 LZF 壓縮字符串  

# 默認都設爲 yes  

# 若是你但願保存子進程節省點 cpu ，你就設置它爲 no ，  

# 不過這個數據集可能就會比較大  

rdbcompression yes  

   

# 是否校驗rdb文件  

rdbchecksum yes  

   

# 設置 dump 的文件位置  

dbfilename dump.rdb  

   

# 工做目錄  

# 例如上面的 dbfilename 只指定了文件名，  

# 可是它會寫入到這個目錄下。這個配置項必定是個目錄，而不能是文件名。  

dir ./  

   

################################# 主從複製 #################################  

   

# 主從複製。使用 slaveof 來讓一個 redis 實例成爲另外一個reids 實例的副本。  

# 注意這個只須要在 slave 上配置。  

#  

# slaveof <masterip> <masterport>  

   

# 若是 master 須要密碼認證，就在這裏設置  

# masterauth <master-password>  

   

# 當一個 slave 與 master 失去聯繫，或者複製正在進行的時候，  

# slave 可能會有兩種表現：  

#  

# 1) 若是爲 yes ，slave 仍然會應答客戶端請求，但返回的數據多是過期，  

#    或者數據多是空的在第一次同步的時候  

#  

# 2) 若是爲 no ，在你執行除了 info he salveof 以外的其餘命令時，  

#    slave 都將返回一個 "SYNC with master in progress" 的錯誤，  

#  

slave-serve-stale-data yes  

   

# 你能夠配置一個 slave 實體是否接受寫入操做。  

# 經過寫入操做來存儲一些短暫的數據對於一個 slave 實例來講多是有用的，  

# 由於相對從 master 從新同步數而言，據數據寫入到 slave 會更容易被刪除。  

# 可是若是客戶端由於一個錯誤的配置寫入，也可能會致使一些問題。  

#  

# 從 redis 2.6 版起，默認 slaves 都是隻讀的。  

#  

# 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.  

# 注意：只讀的 slaves 沒有被設計成在 internet 上暴露給不受信任的客戶端。  

# 它僅僅是一個針對誤用實例的一個保護層。  

slave-read-only yes  

   

# Slaves 在一個預約義的時間間隔內發送 ping 命令到 server 。  

# 你能夠改變這個時間間隔。默認爲 10 秒。  

#  

# repl-ping-slave-period 10  

   

# The following option sets the replication timeout for:  

# 設置主從複製過時時間  

#  

# 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.  

# 這個值必定要比 repl-ping-slave-period 大  

#  

# 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.  

repl-disable-tcp-nodelay no  

   

# 設置主從複製容量大小。這個 backlog 是一個用來在 slaves 被斷開鏈接時  

# 存放 slave 數據的 buffer，因此當一個 slave 想要從新鏈接，一般不但願所有從新同步，  

# 只是部分同步就夠了，僅僅傳遞 slave 在斷開鏈接時丟失的這部分數據。  

#  

# The biggest the replication backlog, the longer the time the slave can be  

# disconnected and later be able to perform a partial resynchronization.  

# 這個值越大，salve 能夠斷開鏈接的時間就越長。  

#  

# The backlog is only allocated once there is at least a slave connected.  

#  

# 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.  

# 在某些時候，master 再也不鏈接 slaves，backlog 將被釋放。  

#  

# A value of 0 means to never release the backlog.  

# 若是設置爲 0 ，意味着毫不釋放 backlog 。  

#  

# repl-backlog-ttl 3600  

   

# 當 master 不能正常工做的時候，Redis Sentinel 會從 slaves 中選出一個新的 master，  

# 這個值越小，就越會被優先選中，可是若是是 0 ， 那是意味着這個 slave 不可能被選中。  

#  

# 默認優先級爲 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:  

#  

# 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.  

   

################################## 安全 ###################################  

   

# 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.  

#   

# 設置認證密碼  

# 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.  

#  

# 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.  

   

################################### 限制 ####################################  

   

# 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).  

#  

# 一旦達到最大限制，redis 將關閉全部的新鏈接  

# 併發送一個‘max number of clients reached’的錯誤。  

#  

# maxclients 10000  

   

# 若是你設置了這個值，當緩存的數據容量達到這個值， redis 將根據你選擇的  

# eviction 策略來移除一些 keys。  

#  

# 若是 redis 不能根據策略移除 keys ，或者是策略被設置爲 ‘noeviction’，  

# redis 將開始響應錯誤給命令，如 set，lpush 等等，  

# 並繼續響應只讀的命令，如 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').  

#  

# 最大使用內存  

# maxmemory <bytes>  

   

# 最大內存策略，你有 5 個選擇。  

#   

# volatile-lru -> remove the key with an expire set using an LRU algorithm  

# volatile-lru -> 使用 LRU 算法移除包含過時設置的 key 。  

# allkeys-lru -> remove any key accordingly to the LRU algorithm  

# allkeys-lru -> 根據 LRU 算法移除全部的 key 。  

# 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  

# noeviction -> 不讓任何 key 過時，只是給寫入操做返回一個錯誤  

#   

# 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:  

#  

# maxmemory-policy noeviction  

   

# LRU and minimal TTL algorithms are not precise algorithms but approximated  

# algorithms (in order to save memory), so you can tune it for speed or  

# accuracy. For default Redis will check five keys and pick the one that was  

# used less recently, you can change the sample size using the following  

# configuration directive.  

#  

# The default of 5 produces good enough results. 10 Approximates very closely  

# true LRU but costs a bit more CPU. 3 is very fast but not very accurate.  

#  

# maxmemory-samples 5  

   

############################## 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.  

   

appendonly no  

   

# The name of the append only file (default: "appendonly.aof")  

   

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".  

   

# 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.  

   

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.  

   

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-time-limit 5000  

   

################################ REDIS 集羣  ###############################  

#  

# 啓用或停用集羣  

# cluster-enabled yes  

   

# Every cluster node has a cluster configuration file. This file is not  

# intended to be edited by hand. It is created and updated by Redis nodes.  

# Every Redis Cluster node requires a different cluster configuration file.  

# Make sure that instances running in the same system does not have  

# overlapping cluster configuration file names.  

#  

# cluster-config-file nodes-6379.conf  

   

# Cluster node timeout is the amount of milliseconds a node must be unreachable   

# for it to be considered in failure state.  

# Most other internal time limits are multiple of the node timeout.  

#  

# cluster-node-timeout 15000  

   

# A slave of a failing master will avoid to start a failover if its data  

# looks too old.  

#  

# There is no simple way for a slave to actually have a exact measure of  

# its "data age", so the following two checks are performed:  

#  

# 1) If there are multiple slaves able to failover, they exchange messages  

#    in order to try to give an advantage to the slave with the best  

#    replication offset (more data from the master processed).  

#    Slaves will try to get their rank by offset, and apply to the start  

#    of the failover a delay proportional to their rank.  

#  

# 2) Every single slave computes the time of the last interaction with  

#    its master. This can be the last ping or command received (if the master  

#    is still in the "connected" state), or the time that elapsed since the  

#    disconnection with the master (if the replication link is currently down).  

#    If the last interaction is too old, the slave will not try to failover  

#    at all.  

#  

# The point "2" can be tuned by user. Specifically a slave will not perform  

# the failover if, since the last interaction with the master, the time  

# elapsed is greater than:  

#  

#   (node-timeout * slave-validity-factor) + repl-ping-slave-period  

#  

# So for example if node-timeout is 30 seconds, and the slave-validity-factor  

# is 10, and assuming a default repl-ping-slave-period of 10 seconds, the  

# slave will not try to failover if it was not able to talk with the master  

# for longer than 310 seconds.  

#  

# A large slave-validity-factor may allow slaves with too old data to failover  

# a master, while a too small value may prevent the cluster from being able to  

# elect a slave at all.  

#  

# For maximum availability, it is possible to set the slave-validity-factor  

# to a value of 0, which means, that slaves will always try to failover the  

# master regardless of the last time they interacted with the master.  

# (However they'll always try to apply a delay proportional to their  

# offset rank).  

#  

# Zero is the only value able to guarantee that when all the partitions heal  

# the cluster will always be able to continue.  

#  

# cluster-slave-validity-factor 10  

   

# Cluster slaves are able to migrate to orphaned masters, that are masters  

# that are left without working slaves. This improves the cluster ability  

# to resist to failures as otherwise an orphaned master can't be failed over  

# in case of failure if it has no working slaves.  

#  

# Slaves migrate to orphaned masters only if there are still at least a  

# given number of other working slaves for their old master. This number  

# is the "migration barrier". A migration barrier of 1 means that a slave  

# will migrate only if there is at least 1 other working slave for its master  

# and so forth. It usually reflects the number of slaves you want for every  

# master in your cluster.  

#  

# Default is 1 (slaves migrate only if their masters remain with at least  

# one slave). To disable migration just set it to a very large value.  

# A value of 0 can be set but is useful only for debugging and dangerous  

# in production.  

#  

# cluster-migration-barrier 1  

   

# In order to setup your cluster make sure to read the documentation  

# available at http://redis.io web site.  

   

################################## 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 128  

   

############################# 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.  

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.  

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-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  

   

# 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.  

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.  

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  

# slave  -> slave clients and MONITOR clients  

# pubsub -> clients subscribed to at least one pubsub channel or pattern  

#  

# The syntax of every client-output-buffer-limit directive is the following:  

#  

# 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.  

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.  

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.  

aof-rewrite-incremental-fsync yes

   你們能夠了解了解。。。