redis配置文件详解（转）

1 # redis 配置文件示例
2
3 # 当你需要为某个配置项指定内存大小的时候，必须要带上单位，
4 # 通常的格式就是 1k 5gb 4m 等酱紫：
5 #
6 # 1k  => 1000 bytes
7 # 1kb => 1024 bytes
8 # 1m  => 1000000 bytes
9 # 1mb => 1024*1024 bytes
10 # 1g  => 1000000000 bytes
11 # 1gb => 1024*1024*1024 bytes
12 #
13 # 单位是不区分大小写的，你写 1K 5GB 4M 也行
14
15 ################################## INCLUDES ###################################
16
17 # 假如说你有一个可用于所有的 redis server 的标准配置模板，
18 # 但针对某些 server 又需要一些个性化的设置，
19 # 你可以使用 include 来包含一些其他的配置文件，这对你来说是非常有用的。
20 #
21 # 但是要注意哦，include 是不能被 config rewrite 命令改写的
22 # 由于 redis 总是以最后的加工线作为一个配置指令值，所以你最好是把 include 放在这个文件的最前面，
23 # 以避免在运行时覆盖配置的改变，相反，你就把它放在后面（外国人真啰嗦）。
24 #
25 # include /path/to/local.conf
26 # include /path/to/other.conf
27
28 ################################ 常用 #####################################
29
30 # 默认情况下 redis 不是作为守护进程运行的，如果你想让它在后台运行，你就把它改成 yes。
31 # 当redis作为守护进程运行的时候，它会写一个 pid 到 /var/run/redis.pid 文件里面。
32 daemonize no
33
34 # 当redis作为守护进程运行的时候，它会把 pid 默认写到 /var/run/redis.pid 文件里面，
35 # 但是你可以在这里自己制定它的文件位置。
36 pidfile /var/run/redis.pid
37
38 # 监听端口号，默认为 6379，如果你设为 0 ，redis 将不在 socket 上监听任何客户端连接。
39 port 6379
40
41 # TCP 监听的最大容纳数量
42 #
43 # 在高并发的环境下，你需要把这个值调高以避免客户端连接缓慢的问题。
44 # Linux 内核会一声不响的把这个值缩小成 /proc/sys/net/core/somaxconn 对应的值，
45 # 所以你要修改这两个值才能达到你的预期。
46 tcp-backlog 511
47
48 # 默认情况下，redis 在 server 上所有有效的网络接口上监听客户端连接。
49 # 你如果只想让它在一个网络接口上监听，那你就绑定一个IP或者多个IP。
50 #
51 # 示例，多个IP用空格隔开:
52 #
53 # bind 192.168.1.100 10.0.0.1
54 # bind 127.0.0.1
55
56 # 指定 unix socket 的路径。
57 #
58 # unixsocket /tmp/redis.sock
59 # unixsocketperm 755
60
61 # 指定在一个 client 空闲多少秒之后关闭连接（0 就是不管它）
62 timeout 0
63
64 # tcp 心跳包。
65 #
66 # 如果设置为非零，则在与客户端缺乏通讯的时候使用 SO_KEEPALIVE 发送 tcp acks 给客户端。
67 # 这个之所有有用，主要由两个原因：
68 #
69 # 1) 防止死的 peers
70 # 2) Take the connection alive from the point of view of network
71 #    equipment in the middle.
72 #
73 # On Linux, the specified value (in seconds) is the period used to send ACKs.
74 # Note that to close the connection the double of the time is needed.
75 # On other kernels the period depends on the kernel configuration.
76 #
77 # A reasonable value for this option is 60 seconds.
78 # 推荐一个合理的值就是60秒
79 tcp-keepalive 0
80
81 # 定义日志级别。
82 # 可以是下面的这些值：
83 # debug (适用于开发或测试阶段)
84 # verbose (many rarely useful info, but not a mess like the debug level)
85 # notice (适用于生产环境)
86 # warning (仅仅一些重要的消息被记录)
87 loglevel notice
88
89 # 指定日志文件的位置
90 logfile ""
91
92 # 要想把日志记录到系统日志，就把它改成 yes，
93 # 也可以可选择性的更新其他的syslog 参数以达到你的要求
94 # syslog-enabled no
95
96 # 设置 syslog 的 identity。
97 # syslog-ident redis
98
99 # 设置 syslog 的 facility，必须是 USER 或者是 LOCAL0-LOCAL7 之间的值。
100 # syslog-facility local0
101
102 # 设置数据库的数目。
103 # 默认数据库是 DB 0，你可以在每个连接上使用 select <dbid> 命令选择一个不同的数据库，
104 # 但是 dbid 必须是一个介于 0 到 databasees - 1 之间的值
105 databases 16
106
107 ################################ 快照 ################################
108 #
109 # 存 DB 到磁盘：
110 #
111 #   格式：save <间隔时间（秒）> <写入次数>
112 #
113 #   根据给定的时间间隔和写入次数将数据保存到磁盘
114 #
115 #   下面的例子的意思是：
116 #   900 秒后如果至少有 1 个 key 的值变化，则保存
117 #   300 秒后如果至少有 10 个 key 的值变化，则保存
118 #   60 秒后如果至少有 10000 个 key 的值变化，则保存
119 #
120 #   注意：你可以注释掉所有的 save 行来停用保存功能。
121 #   也可以直接一个空字符串来实现停用：
122 #   save ""
123
124 save 900 1
125 save 300 10
126 save 60 10000
127
128 # 默认情况下，如果 redis 最后一次的后台保存失败，redis 将停止接受写操作，
129 # 这样以一种强硬的方式让用户知道数据不能正确的持久化到磁盘，
130 # 否则就会没人注意到灾难的发生。
131 #
132 # 如果后台保存进程重新启动工作了，redis 也将自动的允许写操作。
133 #
134 # 然而你要是安装了靠谱的监控，你可能不希望 redis 这样做，那你就改成 no 好了。
135 stop-writes-on-bgsave-error yes
136
137 # 是否在 dump .rdb 数据库的时候使用 LZF 压缩字符串
138 # 默认都设为 yes
139 # 如果你希望保存子进程节省点 cpu ，你就设置它为 no ，
140 # 不过这个数据集可能就会比较大
141 rdbcompression yes
142
143 # 是否校验rdb文件
144 rdbchecksum yes
145
146 # 设置 dump 的文件位置
147 dbfilename dump.rdb
148
149 # 工作目录
150 # 例如上面的 dbfilename 只指定了文件名，
151 # 但是它会写入到这个目录下。这个配置项一定是个目录，而不能是文件名。
152 dir ./
153
154 ################################# 主从复制 #################################
155
156 # 主从复制。使用 slaveof 来让一个 redis 实例成为另一个reids 实例的副本。
157 # 注意这个只需要在 slave 上配置。
158 #
159 # slaveof <masterip> <masterport>
160
161 # 如果 master 需要密码认证，就在这里设置
162 # masterauth <master-password>
163
164 # 当一个 slave 与 master 失去联系，或者复制正在进行的时候，
165 # slave 可能会有两种表现：
166 #
167 # 1) 如果为 yes ，slave 仍然会应答客户端请求，但返回的数据可能是过时，
168 #    或者数据可能是空的在第一次同步的时候
169 #
170 # 2) 如果为 no ，在你执行除了 info he salveof 之外的其他命令时，
171 #    slave 都将返回一个 "SYNC with master in progress" 的错误，
172 #
173 slave-serve-stale-data yes
174
175 # 你可以配置一个 slave 实体是否接受写入操作。
176 # 通过写入操作来存储一些短暂的数据对于一个 slave 实例来说可能是有用的，
177 # 因为相对从 master 重新同步数而言，据数据写入到 slave 会更容易被删除。
178 # 但是如果客户端因为一个错误的配置写入，也可能会导致一些问题。
179 #
180 # 从 redis 2.6 版起，默认 slaves 都是只读的。
181 #
182 # Note: read only slaves are not designed to be exposed to untrusted clients
183 # on the internet. It's just a protection layer against misuse of the instance.
184 # Still a read only slave exports by default all the administrative commands
185 # such as CONFIG, DEBUG, and so forth. To a limited extent you can improve
186 # security of read only slaves using 'rename-command' to shadow all the
187 # administrative / dangerous commands.
188 # 注意：只读的 slaves 没有被设计成在 internet 上暴露给不受信任的客户端。
189 # 它仅仅是一个针对误用实例的一个保护层。
190 slave-read-only yes
191
192 # Slaves 在一个预定义的时间间隔内发送 ping 命令到 server 。
193 # 你可以改变这个时间间隔。默认为 10 秒。
194 #
195 # repl-ping-slave-period 10
196
197 # The following option sets the replication timeout for:
198 # 设置主从复制过期时间
199 #
200 # 1) Bulk transfer I/O during SYNC, from the point of view of slave.
201 # 2) Master timeout from the point of view of slaves (data, pings).
202 # 3) Slave timeout from the point of view of masters (REPLCONF ACK pings).
203 #
204 # It is important to make sure that this value is greater than the value
205 # specified for repl-ping-slave-period otherwise a timeout will be detected
206 # every time there is low traffic between the master and the slave.
207 # 这个值一定要比 repl-ping-slave-period 大
208 #
209 # repl-timeout 60
210
211 # Disable TCP_NODELAY on the slave socket after SYNC?
212 #
213 # If you select "yes" Redis will use a smaller number of TCP packets and
214 # less bandwidth to send data to slaves. But this can add a delay for
215 # the data to appear on the slave side, up to 40 milliseconds with
216 # Linux kernels using a default configuration.
217 #
218 # If you select "no" the delay for data to appear on the slave side will
219 # be reduced but more bandwidth will be used for replication.
220 #
221 # By default we optimize for low latency, but in very high traffic conditions
222 # or when the master and slaves are many hops away, turning this to "yes" may
223 # be a good idea.
224 repl-disable-tcp-nodelay no
225
226 # 设置主从复制容量大小。这个 backlog 是一个用来在 slaves 被断开连接时
227 # 存放 slave 数据的 buffer，所以当一个 slave 想要重新连接，通常不希望全部重新同步，
228 # 只是部分同步就够了，仅仅传递 slave 在断开连接时丢失的这部分数据。
229 #
230 # The biggest the replication backlog, the longer the time the slave can be
231 # disconnected and later be able to perform a partial resynchronization.
232 # 这个值越大，salve 可以断开连接的时间就越长。
233 #
234 # The backlog is only allocated once there is at least a slave connected.
235 #
236 # repl-backlog-size 1mb
237
238 # After a master has no longer connected slaves for some time, the backlog
239 # will be freed. The following option configures the amount of seconds that
240 # need to elapse, starting from the time the last slave disconnected, for
241 # the backlog buffer to be freed.
242 # 在某些时候，master 不再连接 slaves，backlog 将被释放。
243 #
244 # A value of 0 means to never release the backlog.
245 # 如果设置为 0 ，意味着绝不释放 backlog 。
246 #
247 # repl-backlog-ttl 3600
248
249 # 当 master 不能正常工作的时候，Redis Sentinel 会从 slaves 中选出一个新的 master，
250 # 这个值越小，就越会被优先选中，但是如果是 0 ， 那是意味着这个 slave 不可能被选中。
251 #
252 # 默认优先级为 100。
253 slave-priority 100
254
255 # It is possible for a master to stop accepting writes if there are less than
256 # N slaves connected, having a lag less or equal than M seconds.
257 #
258 # The N slaves need to be in "online" state.
259 #
260 # The lag in seconds, that must be <= the specified value, is calculated from
261 # the last ping received from the slave, that is usually sent every second.
262 #
263 # This option does not GUARANTEES that N replicas will accept the write, but
264 # will limit the window of exposure for lost writes in case not enough slaves
265 # are available, to the specified number of seconds.
266 #
267 # For example to require at least 3 slaves with a lag <= 10 seconds use:
268 #
269 # min-slaves-to-write 3
270 # min-slaves-max-lag 10
271 #
272 # Setting one or the other to 0 disables the feature.
273 #
274 # By default min-slaves-to-write is set to 0 (feature disabled) and
275 # min-slaves-max-lag is set to 10.
276
277 ################################## 安全 ###################################
278
279 # Require clients to issue AUTH <PASSWORD> before processing any other
280 # commands.  This might be useful in environments in which you do not trust
281 # others with access to the host running redis-server.
282 #
283 # This should stay commented out for backward compatibility and because most
284 # people do not need auth (e.g. they run their own servers).
285 #
286 # Warning: since Redis is pretty fast an outside user can try up to
287 # 150k passwords per second against a good box. This means that you should
288 # use a very strong password otherwise it will be very easy to break.
289 #
290 # 设置认证密码
291 # requirepass foobared
292
293 # Command renaming.
294 #
295 # It is possible to change the name of dangerous commands in a shared
296 # environment. For instance the CONFIG command may be renamed into something
297 # hard to guess so that it will still be available for internal-use tools
298 # but not available for general clients.
299 #
300 # Example:
301 #
302 # rename-command CONFIG b840fc02d524045429941cc15f59e41cb7be6c52
303 #
304 # It is also possible to completely kill a command by renaming it into
305 # an empty string:
306 #
307 # rename-command CONFIG ""
308 #
309 # Please note that changing the name of commands that are logged into the
310 # AOF file or transmitted to slaves may cause problems.
311
312 ################################### 限制 ####################################
313
314 # Set the max number of connected clients at the same time. By default
315 # this limit is set to 10000 clients, however if the Redis server is not
316 # able to configure the process file limit to allow for the specified limit
317 # the max number of allowed clients is set to the current file limit
318 # minus 32 (as Redis reserves a few file descriptors for internal uses).
319 #
320 # 一旦达到最大限制，redis 将关闭所有的新连接
321 # 并发送一个‘max number of clients reached’的错误。
322 #
323 # maxclients 10000
324
325 # 如果你设置了这个值，当缓存的数据容量达到这个值， redis 将根据你选择的
326 # eviction 策略来移除一些 keys。
327 #
328 # 如果 redis 不能根据策略移除 keys ，或者是策略被设置为 ‘noeviction’，
329 # redis 将开始响应错误给命令，如 set，lpush 等等，
330 # 并继续响应只读的命令，如 get
331 #
332 # This option is usually useful when using Redis as an LRU cache, or to set
333 # a hard memory limit for an instance (using the 'noeviction' policy).
334 #
335 # WARNING: If you have slaves attached to an instance with maxmemory on,
336 # the size of the output buffers needed to feed the slaves are subtracted
337 # from the used memory count, so that network problems / resyncs will
338 # not trigger a loop where keys are evicted, and in turn the output
339 # buffer of slaves is full with DELs of keys evicted triggering the deletion
340 # of more keys, and so forth until the database is completely emptied.
341 #
342 # In short... if you have slaves attached it is suggested that you set a lower
343 # limit for maxmemory so that there is some free RAM on the system for slave
344 # output buffers (but this is not needed if the policy is 'noeviction').
345 #
346 # 最大使用内存
347 # maxmemory <bytes>
348
349 # 最大内存策略，你有 5 个选择。
350 #
351 # volatile-lru -> remove the key with an expire set using an LRU algorithm
352 # volatile-lru -> 使用 LRU 算法移除包含过期设置的 key 。
353 # allkeys-lru -> remove any key accordingly to the LRU algorithm
354 # allkeys-lru -> 根据 LRU 算法移除所有的 key 。
355 # volatile-random -> remove a random key with an expire set
356 # allkeys-random -> remove a random key, any key
357 # volatile-ttl -> remove the key with the nearest expire time (minor TTL)
358 # noeviction -> don't expire at all, just return an error on write operations
359 # noeviction -> 不让任何 key 过期，只是给写入操作返回一个错误
360 #
361 # Note: with any of the above policies, Redis will return an error on write
362 #       operations, when there are not suitable keys for eviction.
363 #
364 #       At the date of writing this commands are: set setnx setex append
365 #       incr decr rpush lpush rpushx lpushx linsert lset rpoplpush sadd
366 #       sinter sinterstore sunion sunionstore sdiff sdiffstore zadd zincrby
367 #       zunionstore zinterstore hset hsetnx hmset hincrby incrby decrby
368 #       getset mset msetnx exec sort
369 #
370 # The default is:
371 #
372 # maxmemory-policy noeviction
373
374 # LRU and minimal TTL algorithms are not precise algorithms but approximated
375 # algorithms (in order to save memory), so you can tune it for speed or
376 # accuracy. For default Redis will check five keys and pick the one that was
377 # used less recently, you can change the sample size using the following
378 # configuration directive.
379 #
380 # The default of 5 produces good enough results. 10 Approximates very closely
381 # true LRU but costs a bit more CPU. 3 is very fast but not very accurate.
382 #
383 # maxmemory-samples 5
384
385 ############################## APPEND ONLY MODE ###############################
386
387 # By default Redis asynchronously dumps the dataset on disk. This mode is
388 # good enough in many applications, but an issue with the Redis process or
389 # a power outage may result into a few minutes of writes lost (depending on
390 # the configured save points).
391 #
392 # The Append Only File is an alternative persistence mode that provides
393 # much better durability. For instance using the default data fsync policy
394 # (see later in the config file) Redis can lose just one second of writes in a
395 # dramatic event like a server power outage, or a single write if something
396 # wrong with the Redis process itself happens, but the operating system is
397 # still running correctly.
398 #
399 # AOF and RDB persistence can be enabled at the same time without problems.
400 # If the AOF is enabled on startup Redis will load the AOF, that is the file
401 # with the better durability guarantees.
402 #
403 # Please check http://redis.io/topics/persistence for more information.
404
405 appendonly no
406
407 # The name of the append only file (default: "appendonly.aof")
408
409 appendfilename "appendonly.aof"
410
411 # The fsync() call tells the Operating System to actually write data on disk
412 # instead to wait for more data in the output buffer. Some OS will really flush
413 # data on disk, some other OS will just try to do it ASAP.
414 #
415 # Redis supports three different modes:
416 #
417 # no: don't fsync, just let the OS flush the data when it wants. Faster.
418 # always: fsync after every write to the append only log . Slow, Safest.
419 # everysec: fsync only one time every second. Compromise.
420 #
421 # The default is "everysec", as that's usually the right compromise between
422 # speed and data safety. It's up to you to understand if you can relax this to
423 # "no" that will let the operating system flush the output buffer when
424 # it wants, for better performances (but if you can live with the idea of
425 # some data loss consider the default persistence mode that's snapshotting),
426 # or on the contrary, use "always" that's very slow but a bit safer than
427 # everysec.
428 #
429 # More details please check the following article:
430 # http://antirez.com/post/redis-persistence-demystified.html 431 #
432 # If unsure, use "everysec".
433
434 # appendfsync always
435 appendfsync everysec
436 # appendfsync no
437
438 # When the AOF fsync policy is set to always or everysec, and a background
439 # saving process (a background save or AOF log background rewriting) is
440 # performing a lot of I/O against the disk, in some Linux configurations
441 # Redis may block too long on the fsync() call. Note that there is no fix for
442 # this currently, as even performing fsync in a different thread will block
443 # our synchronous write(2) call.
444 #
445 # In order to mitigate this problem it's possible to use the following option
446 # that will prevent fsync() from being called in the main process while a
447 # BGSAVE or BGREWRITEAOF is in progress.
448 #
449 # This means that while another child is saving, the durability of Redis is
450 # the same as "appendfsync none". In practical terms, this means that it is
451 # possible to lose up to 30 seconds of log in the worst scenario (with the
452 # default Linux settings).
453 #
454 # If you have latency problems turn this to "yes". Otherwise leave it as
455 # "no" that is the safest pick from the point of view of durability.
456
457 no-appendfsync-on-rewrite no
458
459 # Automatic rewrite of the append only file.
460 # Redis is able to automatically rewrite the log file implicitly calling
461 # BGREWRITEAOF when the AOF log size grows by the specified percentage.
462 #
463 # This is how it works: Redis remembers the size of the AOF file after the
464 # latest rewrite (if no rewrite has happened since the restart, the size of
465 # the AOF at startup is used).
466 #
467 # This base size is compared to the current size. If the current size is
468 # bigger than the specified percentage, the rewrite is triggered. Also
469 # you need to specify a minimal size for the AOF file to be rewritten, this
470 # is useful to avoid rewriting the AOF file even if the percentage increase
471 # is reached but it is still pretty small.
472 #
473 # Specify a percentage of zero in order to disable the automatic AOF
474 # rewrite feature.
475
476 auto-aof-rewrite-percentage 100
477 auto-aof-rewrite-min-size 64mb
478
479 ################################ LUA SCRIPTING  ###############################
480
481 # Max execution time of a Lua script in milliseconds.
482 #
483 # If the maximum execution time is reached Redis will log that a script is
484 # still in execution after the maximum allowed time and will start to
485 # reply to queries with an error.
486 #
487 # When a long running script exceed the maximum execution time only the
488 # SCRIPT KILL and SHUTDOWN NOSAVE commands are available. The first can be
489 # used to stop a script that did not yet called write commands. The second
490 # is the only way to shut down the server in the case a write commands was
491 # already issue by the script but the user don't want to wait for the natural
492 # termination of the script.
493 #
494 # Set it to 0 or a negative value for unlimited execution without warnings.
495 lua-time-limit 5000
496
497 ################################ REDIS 集群  ###############################
498 #
499 # 启用或停用集群
500 # cluster-enabled yes
501
502 # Every cluster node has a cluster configuration file. This file is not
503 # intended to be edited by hand. It is created and updated by Redis nodes.
504 # Every Redis Cluster node requires a different cluster configuration file.
505 # Make sure that instances running in the same system does not have
506 # overlapping cluster configuration file names.
507 #
508 # cluster-config-file nodes-6379.conf
509
510 # Cluster node timeout is the amount of milliseconds a node must be unreachable
511 # for it to be considered in failure state.
512 # Most other internal time limits are multiple of the node timeout.
513 #
514 # cluster-node-timeout 15000
515
516 # A slave of a failing master will avoid to start a failover if its data
517 # looks too old.
518 #
519 # There is no simple way for a slave to actually have a exact measure of
520 # its "data age", so the following two checks are performed:
521 #
522 # 1) If there are multiple slaves able to failover, they exchange messages
523 #    in order to try to give an advantage to the slave with the best
524 #    replication offset (more data from the master processed).
525 #    Slaves will try to get their rank by offset, and apply to the start
526 #    of the failover a delay proportional to their rank.
527 #
528 # 2) Every single slave computes the time of the last interaction with
529 #    its master. This can be the last ping or command received (if the master
530 #    is still in the "connected" state), or the time that elapsed since the
531 #    disconnection with the master (if the replication link is currently down).
532 #    If the last interaction is too old, the slave will not try to failover
533 #    at all.
534 #
535 # The point "2" can be tuned by user. Specifically a slave will not perform
536 # the failover if, since the last interaction with the master, the time
537 # elapsed is greater than:
538 #
539 #   (node-timeout * slave-validity-factor) + repl-ping-slave-period
540 #
541 # So for example if node-timeout is 30 seconds, and the slave-validity-factor
542 # is 10, and assuming a default repl-ping-slave-period of 10 seconds, the
543 # slave will not try to failover if it was not able to talk with the master
544 # for longer than 310 seconds.
545 #
546 # A large slave-validity-factor may allow slaves with too old data to failover
547 # a master, while a too small value may prevent the cluster from being able to
548 # elect a slave at all.
549 #
550 # For maximum availability, it is possible to set the slave-validity-factor
551 # to a value of 0, which means, that slaves will always try to failover the
552 # master regardless of the last time they interacted with the master.
553 # (However they'll always try to apply a delay proportional to their
554 # offset rank).
555 #
556 # Zero is the only value able to guarantee that when all the partitions heal
557 # the cluster will always be able to continue.
558 #
559 # cluster-slave-validity-factor 10
560
561 # Cluster slaves are able to migrate to orphaned masters, that are masters
562 # that are left without working slaves. This improves the cluster ability
563 # to resist to failures as otherwise an orphaned master can't be failed over
564 # in case of failure if it has no working slaves.
565 #
566 # Slaves migrate to orphaned masters only if there are still at least a
567 # given number of other working slaves for their old master. This number
568 # is the "migration barrier". A migration barrier of 1 means that a slave
569 # will migrate only if there is at least 1 other working slave for its master
570 # and so forth. It usually reflects the number of slaves you want for every
571 # master in your cluster.
572 #
573 # Default is 1 (slaves migrate only if their masters remain with at least
574 # one slave). To disable migration just set it to a very large value.
575 # A value of 0 can be set but is useful only for debugging and dangerous
576 # in production.
577 #
578 # cluster-migration-barrier 1
579
580 # In order to setup your cluster make sure to read the documentation
581 # available at http://redis.io web site.
582
583 ################################## SLOW LOG ###################################
584
585 # The Redis Slow Log is a system to log queries that exceeded a specified
586 # execution time. The execution time does not include the I/O operations
587 # like talking with the client, sending the reply and so forth,
588 # but just the time needed to actually execute the command (this is the only
589 # stage of command execution where the thread is blocked and can not serve
590 # other requests in the meantime).
591 #
592 # You can configure the slow log with two parameters: one tells Redis
593 # what is the execution time, in microseconds, to exceed in order for the
594 # command to get logged, and the other parameter is the length of the
595 # slow log. When a new command is logged the oldest one is removed from the
596 # queue of logged commands.
597
598 # The following time is expressed in microseconds, so 1000000 is equivalent
599 # to one second. Note that a negative number disables the slow log, while
600 # a value of zero forces the logging of every command.
601 slowlog-log-slower-than 10000
602
603 # There is no limit to this length. Just be aware that it will consume memory.
604 # You can reclaim memory used by the slow log with SLOWLOG RESET.
605 slowlog-max-len 128
606
607 ############################# Event notification ##############################
608
609 # Redis can notify Pub/Sub clients about events happening in the key space.
610 # This feature is documented at http://redis.io/topics/keyspace-events 611 #
612 # For instance if keyspace events notification is enabled, and a client
613 # performs a DEL operation on key "foo" stored in the Database 0, two
614 # messages will be published via Pub/Sub:
615 #
616 # PUBLISH __keyspace@0__:foo del
617 # PUBLISH __keyevent@0__:del foo
618 #
619 # It is possible to select the events that Redis will notify among a set
620 # of classes. Every class is identified by a single character:
621 #
622 #  K     Keyspace events, published with __keyspace@<db>__ prefix.
623 #  E     Keyevent events, published with __keyevent@<db>__ prefix.
624 #  g     Generic commands (non-type specific) like DEL, EXPIRE, RENAME, ...
625 #  $     String commands
626 #  l     List commands
627 #  s     Set commands
628 #  h     Hash commands
629 #  z     Sorted set commands
630 #  x     Expired events (events generated every time a key expires)
631 #  e     Evicted events (events generated when a key is evicted for maxmemory)
632 #  A     Alias for g$lshzxe, so that the "AKE" string means all the events.
633 #
634 #  The "notify-keyspace-events" takes as argument a string that is composed
635 #  by zero or multiple characters. The empty string means that notifications
636 #  are disabled at all.
637 #
638 #  Example: to enable list and generic events, from the point of view of the
639 #           event name, use:
640 #
641 #  notify-keyspace-events Elg
642 #
643 #  Example 2: to get the stream of the expired keys subscribing to channel
644 #             name __keyevent@0__:expired use:
645 #
646 #  notify-keyspace-events Ex
647 #
648 #  By default all notifications are disabled because most users don't need
649 #  this feature and the feature has some overhead. Note that if you don't
650 #  specify at least one of K or E, no events will be delivered.
651 notify-keyspace-events ""
652
653 ############################### ADVANCED CONFIG ###############################
654
655 # Hashes are encoded using a memory efficient data structure when they have a
656 # small number of entries, and the biggest entry does not exceed a given
657 # threshold. These thresholds can be configured using the following directives.
658 hash-max-ziplist-entries 512
659 hash-max-ziplist-value 64
660
661 # Similarly to hashes, small lists are also encoded in a special way in order
662 # to save a lot of space. The special representation is only used when
663 # you are under the following limits:
664 list-max-ziplist-entries 512
665 list-max-ziplist-value 64
666
667 # Sets have a special encoding in just one case: when a set is composed
668 # of just strings that happens to be integers in radix 10 in the range
669 # of 64 bit signed integers.
670 # The following configuration setting sets the limit in the size of the
671 # set in order to use this special memory saving encoding.
672 set-max-intset-entries 512
673
674 # Similarly to hashes and lists, sorted sets are also specially encoded in
675 # order to save a lot of space. This encoding is only used when the length and
676 # elements of a sorted set are below the following limits:
677 zset-max-ziplist-entries 128
678 zset-max-ziplist-value 64
679
680 # HyperLogLog sparse representation bytes limit. The limit includes the
681 # 16 bytes header. When an HyperLogLog using the sparse representation crosses
682 # this limit, it is converted into the dense representation.
683 #
684 # A value greater than 16000 is totally useless, since at that point the
685 # dense representation is more memory efficient.
686 #
687 # The suggested value is ~ 3000 in order to have the benefits of
688 # the space efficient encoding without slowing down too much PFADD,
689 # which is O(N) with the sparse encoding. The value can be raised to
690 # ~ 10000 when CPU is not a concern, but space is, and the data set is
691 # composed of many HyperLogLogs with cardinality in the 0 - 15000 range.
692 hll-sparse-max-bytes 3000
693
694 # Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in
695 # order to help rehashing the main Redis hash table (the one mapping top-level
696 # keys to values). The hash table implementation Redis uses (see dict.c)
697 # performs a lazy rehashing: the more operation you run into a hash table
698 # that is rehashing, the more rehashing "steps" are performed, so if the
699 # server is idle the rehashing is never complete and some more memory is used
700 # by the hash table.
701 #
702 # The default is to use this millisecond 10 times every second in order to
703 # active rehashing the main dictionaries, freeing memory when possible.
704 #
705 # If unsure:
706 # use "activerehashing no" if you have hard latency requirements and it is
707 # not a good thing in your environment that Redis can reply form time to time
708 # to queries with 2 milliseconds delay.
709 #
710 # use "activerehashing yes" if you don't have such hard requirements but
711 # want to free memory asap when possible.
712 activerehashing yes
713
714 # The client output buffer limits can be used to force disconnection of clients
715 # that are not reading data from the server fast enough for some reason (a
716 # common reason is that a Pub/Sub client can't consume messages as fast as the
717 # publisher can produce them).
718 #
719 # The limit can be set differently for the three different classes of clients:
720 #
721 # normal -> normal clients
722 # slave  -> slave clients and MONITOR clients
723 # pubsub -> clients subscribed to at least one pubsub channel or pattern
724 #
725 # The syntax of every client-output-buffer-limit directive is the following:
726 #
727 # client-output-buffer-limit <class> <hard limit> <soft limit> <soft seconds>
728 #
729 # A client is immediately disconnected once the hard limit is reached, or if
730 # the soft limit is reached and remains reached for the specified number of
731 # seconds (continuously).
732 # So for instance if the hard limit is 32 megabytes and the soft limit is
733 # 16 megabytes / 10 seconds, the client will get disconnected immediately
734 # if the size of the output buffers reach 32 megabytes, but will also get
735 # disconnected if the client reaches 16 megabytes and continuously overcomes
736 # the limit for 10 seconds.
737 #
738 # By default normal clients are not limited because they don't receive data
739 # without asking (in a push way), but just after a request, so only
740 # asynchronous clients may create a scenario where data is requested faster
741 # than it can read.
742 #
743 # Instead there is a default limit for pubsub and slave clients, since
744 # subscribers and slaves receive data in a push fashion.
745 #
746 # Both the hard or the soft limit can be disabled by setting them to zero.
747 client-output-buffer-limit normal 0 0 0
748 client-output-buffer-limit slave 256mb 64mb 60
749 client-output-buffer-limit pubsub 32mb 8mb 60
750
751 # Redis calls an internal function to perform many background tasks, like
752 # closing connections of clients in timeout, purging expired keys that are
753 # never requested, and so forth.
754 #
755 # Not all tasks are performed with the same frequency, but Redis checks for
756 # tasks to perform accordingly to the specified "hz" value.
757 #
758 # By default "hz" is set to 10. Raising the value will use more CPU when
759 # Redis is idle, but at the same time will make Redis more responsive when
760 # there are many keys expiring at the same time, and timeouts may be
761 # handled with more precision.
762 #
763 # The range is between 1 and 500, however a value over 100 is usually not
764 # a good idea. Most users should use the default of 10 and raise this up to
765 # 100 only in environments where very low latency is required.
766 hz 10
767
768 # When a child rewrites the AOF file, if the following option is enabled
769 # the file will be fsync-ed every 32 MB of data generated. This is useful
770 # in order to commit the file to the disk more incrementally and avoid
771 # big latency spikes.
772 aof-rewrite-incremental-fsync yes