一、units单位
1 # Redis configuration file example.
2 #
3 # Note that in order to read the configuration file, Redis must be
4 # started with the file path as first argument:
5 #
6 # ./redis-server /path/to/redis.conf
7
8 # Note on units: when memory size is needed, it is possible to specify
9 # it in the usual form of 1k 5GB 4M and so forth:
10 #
11 # 1k => 1000 bytes
12 # 1kb => 1024 bytes
13 # 1m => 1000000 bytes
14 # 1mb => 1024*1024 bytes
15 # 1g => 1000000000 bytes
16 # 1gb => 1024*1024*1024 bytes
17 #
18 # units are case insensitive so 1GB 1Gb 1gB are all the same.
19
1. 配置大小单位,开头定义了一些基本的度量单位，只支持bytes，不支持bit 2. 对大小写不敏感
二、INCLUDES 20 ################################## INCLUDES ###################################
21
22 # Include one or more other config files here. This is useful if you
23 # have a standard template that goes to all Redis servers but also need
24 # to customize a few per-server settings. Include files can include
25 # other files, so use this wisely.
26 #
27 # Notice option "include" won't be rewritten by command "CONFIG REWRITE"
28 # from admin or Redis Sentinel. Since Redis always uses the last processed
29 # line as value of a configuration directive, you'd better put includes
30 # at the beginning of this file to avoid overwriting config change at runtime.
31 #
32 # If instead you are interested in using includes to override configuration
33 # options, it is better to use include as the last line.
34 #
35 # include /path/to/local.conf
36 # include /path/to/other.conf
37
1. 和我们的Struts2配置文件类似，可以通过includes包含，redis.conf可以作为总闸，包含其他
三、NETWORK 38 ################################## NETWORK #####################################
39
40 # By default, if no "bind" configuration directive is specified, Redis listens
41 # for connections from all the network interfaces available on the server.
42 # It is possible to listen to just one or multiple selected interfaces using
43 # the "bind" configuration directive, followed by one or more IP addresses.
44 #
45 # Examples:
46 #
47 # bind 192.168.1.100 10.0.0.1
48 # bind 127.0.0.1 ::1
49 #
50 # ~~~ WARNING ~~~ If the computer running Redis is directly exposed to the
51 # internet, binding to all the interfaces is dangerous and will expose the
52 # instance to everybody on the internet. So by default we uncomment the
53 # following bind directive, that will force Redis to listen only into
54 # the IPv4 lookback interface address (this means Redis will be able to
55 # accept connections only from clients running into the same computer it
56 # is running).
57 #
58 # IF YOU ARE SURE YOU WANT YOUR INSTANCE TO LISTEN TO ALL THE INTERFACES
59 # JUST COMMENT THE FOLLOWING LINE.
60 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
61 bind 127.0.0.1
62
63 # Protected mode is a layer of security protection, in order to avoid that
64 # Redis instances left open on the internet are accessed and exploited.
65 #
66 # When protected mode is on and if:
67 #
68 # 1) The server is not binding explicitly to a set of addresses using the
69 # "bind" directive.
70 # 2) No password is configured.
71 #
72 # The server only accepts connections from clients connecting from the
73 # IPv4 and IPv6 loopback addresses 127.0.0.1 and ::1, and from Unix domain
74 # sockets.
75 #
76 # By default protected mode is enabled. You should disable it only if
77 # you are sure you want clients from other hosts to connect to Redis
78 # even if no authentication is configured, nor a specific set of interfaces
79 # are explicitly listed using the "bind" directive.
80 protected-mode no
81
82 # Accept connections on the specified port, default is 6379 (IANA #815344).
83 # If port 0 is specified Redis will not listen on a TCP socket.
84 port 6379
85
86 # TCP listen() backlog.
87 #
88 # In high requests-per-second environments you need an high backlog in order
89 # to avoid slow clients connections issues. Note that the Linux kernel
90 # will silently truncate it to the value of /proc/sys/net/core/somaxconn so
91 # make sure to raise both the value of somaxconn and tcp_max_syn_backlog
92 # in order to get the desired effect.
93 tcp-backlog 511
94
95 # Unix socket.
96 #
97 # Specify the path for the Unix socket that will be used to listen for
98 # incoming connections. There is no default, so Redis will not listen
99 # on a unix socket when not specified.
100 #
101 # unixsocket /tmp/redis.sock
102 # unixsocketperm 700
103
104 # Close the connection after a client is idle for N seconds (0 to disable)
105 timeout 0
106
107 # TCP keepalive.
108 #
109 # If non-zero, use SO_KEEPALIVE to send TCP ACKs to clients in absence
110 # of communication. This is useful for two reasons:
111 #
112 # 1) Detect dead peers.
113 # 2) Take the connection alive from the point of view of network
114 # equipment in the middle.
115 #
116 # On Linux, the specified value (in seconds) is the period used to send ACKs.
117 # Note that to close the connection the double of the time is needed.
118 # On other kernels the period depends on the kernel configuration.
119 #
120 # A reasonable value for this option is 300 seconds, which is the new
121 # Redis default starting with Redis 3.2.1.
122 tcp-keepalive 300
123
1. bind 绑定的主机地址，你可以绑定单一接口，如果没有绑定，所有接口都会监听到来的连接。 2. protected-mode 设置外部网络连接redis服务，设置no，外部网络可以直接访问；设置yes，需配置bind ip或者设置访问密码。 3. port 指定Redis监听端口，默认端口为6379，如果指定0端口，表示Redis不监听TCP连接。 4. tcp-backlog 设置tcp的backlog，backlog其实是一个连接队列，backlog队列总和=未完成三次握手队列 + 已经完成三次握手队列。在高并发环境下你需要一个高backlog值来避免慢客户端连接问题。注意Linux内核会将这个值减小到/proc/sys/net/core/somaxconn的值，所以需要确认增大somaxconn和tcp_max_syn_backlog两个值来达到想要的效果。 5. timeout 当客户端闲置多长时间后关闭连接，如果指定为0，表示关闭该功能。 6. tcp-keepalive 单位是秒，表示将周期性的使用SO_KEEPALIVE检测客户端是否还处于健康状态，避免服务器一直阻塞，官方给出的建议值是300s，如果设置为0，则不会周期性的检测。
四、GENERAL 124 ################################# GENERAL #####################################
125
126 # By default Redis does not run as a daemon. Use 'yes' if you need it.
127 # Note that Redis will write a pid file in /var/run/redis.pid when daemonized.
128 daemonize no
129
130 # If you run Redis from upstart or systemd, Redis can interact with your
131 # supervision tree. Options:
132 # supervised no - no supervision interaction
133 # supervised upstart - signal upstart by putting Redis into SIGSTOP mode
134 # supervised systemd - signal systemd by writing READY=1 to $NOTIFY_SOCKET
135 # supervised auto - detect upstart or systemd method based on
136 # UPSTART_JOB or NOTIFY_SOCKET environment variables
137 # Note: these supervision methods only signal "process is ready."
138 # They do not enable continuous liveness pings back to your supervisor.
139 supervised no
140
141 # If a pid file is specified, Redis writes it where specified at startup
142 # and removes it at exit.
143 #
144 # When the server runs non daemonized, no pid file is created if none is
145 # specified in the configuration. When the server is daemonized, the pid file
146 # is used even if not specified, defaulting to "/var/run/redis.pid".
147 #
148 # Creating a pid file is best effort: if Redis is not able to create it
149 # nothing bad happens, the server will start and run normally.
150 pidfile /var/run/redis_6379.pid
151
152 # Specify the server verbosity level.
153 # This can be one of:
154 # debug (a lot of information, useful for development/testing)
155 # verbose (many rarely useful info, but not a mess like the debug level)
156 # notice (moderately verbose, what you want in production probably)
157 # warning (only very important / critical messages are logged)
158 loglevel notice
159
160 # Specify the log file name. Also the empty string can be used to force
161 # Redis to log on the standard output. Note that if you use standard
162 # output for logging but daemonize, logs will be sent to /dev/null
163 logfile ""
164
165 # To enable logging to the system logger, just set 'syslog-enabled' to yes,
166 # and optionally update the other syslog parameters to suit your needs.
167 # syslog-enabled no
168
169 # Specify the syslog identity.
170 # syslog-ident redis
171
172 # Specify the syslog facility. Must be USER or between LOCAL0-LOCAL7.
173 # syslog-facility local0
174
175 # Set the number of databases. The default database is DB 0, you can select
176 # a different one on a per-connection basis using SELECT where
177 # dbid is a number between 0 and 'databases'-1
178 databases 16
179
1. daemonize 默认值为 no，默认不是以守护进程的方式运行，可以通过该配置项修改，使用yes启用守护进程 2. pidfile 当Redis以守护进程方式运行时，Redis默认会把pid写入/var/run/redis.pid文件，可以通过pidfile指定 3. loglevel 指定日志记录级别，Redis总共支持四个级别：debug、verbose、notice、warning，默认为verbose 4. logfile 日志记录方式，默认为标准输出，如果配置Redis为守护进程方式运行，而这里又配置为日志记录方式为标准输出，则日志将会发送给/dev/null。 5. databases 设置数据库的数量，默认数据库为0，可以使用SELECT 命令在连接上指定数据库id
五、SNAPSHOTTING180 ################################ SNAPSHOTTING ################################
181 #
182 # Save the DB on disk:
183 #
184 # save
185 #
186 # Will save the DB if both the given number of seconds and the given
187 # number of write operations against the DB occurred.
188 #
189 # In the example below the behaviour will be to save:
190 # after 900 sec (15 min) if at least 1 key changed
191 # after 300 sec (5 min) if at least 10 keys changed
192 # after 60 sec if at least 10000 keys changed
193 #
194 # Note: you can disable saving completely by commenting out all "save" lines.
195 #
196 # It is also possible to remove all the previously configured save
197 # points by adding a save directive with a single empty string argument
198 # like in the following example:
199 #
200 # save ""
201
202 save 900 1
203 save 300 10
204 save 60 10000
205
206 # By default Redis will stop accepting writes if RDB snapshots are enabled
207 # (at least one save point) and the latest background save failed.
208 # This will make the user aware (in a hard way) that data is not persisting
209 # on disk properly, otherwise chances are that no one will notice and some
210 # disaster will happen.
211 #
212 # If the background saving process will start working again Redis will
213 # automatically allow writes again.
214 #
215 # However if you have setup your proper monitoring of the Redis server
216 # and persistence, you may want to disable this feature so that Redis will
217 # continue to work as usual even if there are problems with disk,
218 # permissions, and so forth.
219 stop-writes-on-bgsave-error yes
220
221 # Compress string objects using LZF when dump .rdb databases?
222 # For default that's set to 'yes' as it's almost always a win.
223 # If you want to save some CPU in the saving child set it to 'no' but
224 # the dataset will likely be bigger if you have compressible values or keys.
225 rdbcompression yes
226
227 # Since version 5 of RDB a CRC64 checksum is placed at the end of the file.
228 # This makes the format more resistant to corruption but there is a performance
229 # hit to pay (around 10%) when saving and loading RDB files, so you can disable it
230 # for maximum performances.
231 #
232 # RDB files created with checksum disabled have a checksum of zero that will
233 # tell the loading code to skip the check.
234 rdbchecksum yes
235
236 # The filename where to dump the DB
237 dbfilename dump.rdb
238
239 # The working directory.
240 #
241 # The DB will be written inside this directory, with the filename specified
242 # above using the 'dbfilename' configuration directive.
243 #
244 # The Append Only File will also be created inside this directory.
245 #
246 # Note that you must specify a directory here, not a file name.
247 dir ./
248
- 数据保存频率： 1、save 900 1 900秒后保存，至少有1个key被更改时才会触发 2、save 300 10 300秒后保存，至少有10个key被更改时才会触发 3、save 60 10000 60秒后保存，至少有10000个key被更改时才会触发
- stop-writes-on-bgsave-error yes 最近一次save操作失败则停止写操作
- rdbcompression yes 启用压缩
- rdbchecksum yes 启用CRC64校验码，当然这个会影响一部份性能
- dbfilename dump.rdb 指定存储数据的文件名
- dir ./ 指定工作目录，rdb文件和aof文件都会存放在这个目录中，默认为当前目录
249 ################################# REPLICATION #################################
250
251 # Master-Slave replication. Use slaveof to make a Redis instance a copy of
252 # another Redis server. A few things to understand ASAP about Redis replication.
253 #
254 # 1) Redis replication is asynchronous, but you can configure a master to
255 # stop accepting writes if it appears to be not connected with at least
256 # a given number of slaves.
257 # 2) Redis slaves are able to perform a partial resynchronization with the
258 # master if the replication link is lost for a relatively small amount of
259 # time. You may want to configure the replication backlog size (see the next
260 # sections of this file) with a sensible value depending on your needs.
261 # 3) Replication is automatic and does not need user intervention. After a
262 # network partition slaves automatically try to reconnect to masters
263 # and resynchronize with them.
264 #
265 # slaveof
266
267 # If the master is password protected (using the "requirepass" configuration
268 # directive below) it is possible to tell the slave to authenticate before
269 # starting the replication synchronization process, otherwise the master will
270 # refuse the slave request.
271 #
272 # masterauth
273
274 # When a slave loses its connection with the master, or when the replication
275 # is still in progress, the slave can act in two different ways:
276 #
277 # 1) if slave-serve-stale-data is set to 'yes' (the default) the slave will
278 # still reply to client requests, possibly with out of date data, or the
279 # data set may just be empty if this is the first synchronization.
280 #
281 # 2) if slave-serve-stale-data is set to 'no' the slave will reply with
282 # an error "SYNC with master in progress" to all the kind of commands
283 # but to INFO and SLAVEOF.
284 #
285 slave-serve-stale-data yes
286
287 # You can configure a slave instance to accept writes or not. Writing against
288 # a slave instance may be useful to store some ephemeral data (because data
289 # written on a slave will be easily deleted after resync with the master) but
290 # may also cause problems if clients are writing to it because of a
291 # misconfiguration.
292 #
293 # Since Redis 2.6 by default slaves are read-only.
294 #
295 # Note: read only slaves are not designed to be exposed to untrusted clients
296 # on the internet. It's just a protection layer against misuse of the instance.
297 # Still a read only slave exports by default all the administrative commands
298 # such as CONFIG, DEBUG, and so forth. To a limited extent you can improve
299 # security of read only slaves using 'rename-command' to shadow all the
300 # administrative / dangerous commands.
301 slave-read-only yes
302
303 # Replication SYNC strategy: disk or socket.
304 #
305 # -------------------------------------------------------
306 # WARNING: DISKLESS REPLICATION IS EXPERIMENTAL CURRENTLY
307 # -------------------------------------------------------
308 #
309 # New slaves and reconnecting slaves that are not able to continue the replication
310 # process just receiving differences, need to do what is called a "full
311 # synchronization". An RDB file is transmitted from the master to the slaves.
312 # The transmission can happen in two different ways:
313 #
314 # 1) Disk-backed: The Redis master creates a new process that writes the RDB
315 # file on disk. Later the file is transferred by the parent
316 # process to the slaves incrementally.
317 # 2) Diskless: The Redis master creates a new process that directly writes the
318 # RDB file to slave sockets, without touching the disk at all.
319 #
320 # With disk-backed replication, while the RDB file is generated, more slaves
321 # can be queued and served with the RDB file as soon as the current child producing
322 # the RDB file finishes its work. With diskless replication instead once
323 # the transfer starts, new slaves arriving will be queued and a new transfer
324 # will start when the current one terminates.
325 #
326 # When diskless replication is used, the master waits a configurable amount of
327 # time (in seconds) before starting the transfer in the hope that multiple slaves
328 # will arrive and the transfer can be parallelized.
329 #
330 # With slow disks and fast (large bandwidth) networks, diskless replication
331 # works better.
332 repl-diskless-sync no
333
334 # When diskless replication is enabled, it is possible to configure the delay
335 # the server waits in order to spawn the child that transfers the RDB via socket
336 # to the slaves.
337 #
338 # This is important since once the transfer starts, it is not possible to serve
339 # new slaves arriving, that will be queued for the next RDB transfer, so the serve r
340 # waits a delay in order to let more slaves arrive.
341 #
342 # The delay is specified in seconds, and by default is 5 seconds. To disable
343 # it entirely just set it to 0 seconds and the transfer will start ASAP.
344 repl-diskless-sync-delay 5
345
346 # Slaves send PINGs to server in a predefined interval. It's possible to change
347 # this interval with the repl_ping_slave_period option. The default value is 10
348 # seconds.
349 #
350 # repl-ping-slave-period 10
351
352 # The following option sets the replication timeout for:
353 #
354 # 1) Bulk transfer I/O during SYNC, from the point of view of slave.
355 # 2) Master timeout from the point of view of slaves (data, pings).
356 # 3) Slave timeout from the point of view of masters (REPLCONF ACK pings).
357 #
358 # It is important to make sure that this value is greater than the value
359 # specified for repl-ping-slave-period otherwise a timeout will be detected
360 # every time there is low traffic between the master and the slave.
361 #
362 # repl-timeout 60
363
364 # Disable TCP_NODELAY on the slave socket after SYNC?
365 #
366 # If you select "yes" Redis will use a smaller number of TCP packets and
367 # less bandwidth to send data to slaves. But this can add a delay for
368 # the data to appear on the slave side, up to 40 milliseconds with
369 # Linux kernels using a default configuration.
370 #
371 # If you select "no" the delay for data to appear on the slave side will
372 # be reduced but more bandwidth will be used for replication.
373 #
374 # By default we optimize for low latency, but in very high traffic conditions
375 # or when the master and slaves are many hops away, turning this to "yes" may
376 # be a good idea.
377 repl-disable-tcp-nodelay no
378
379 # Set the replication backlog size. The backlog is a buffer that accumulates
380 # slave data when slaves are disconnected for some time, so that when a slave
381 # wants to reconnect again, often a full resync is not needed, but a partial
382 # resync is enough, just passing the portion of data the slave missed while
383 # disconnected.
384 #
385 # The bigger the replication backlog, the longer the time the slave can be
386 # disconnected and later be able to perform a partial resynchronization.
387 #
388 # The backlog is only allocated once there is at least a slave connected.
389 #
390 # repl-backlog-size 1mb
391
392 # After a master has no longer connected slaves for some time, the backlog
393 # will be freed. The following option configures the amount of seconds that
394 # need to elapse, starting from the time the last slave disconnected, for
395 # the backlog buffer to be freed.
396 #
397 # A value of 0 means to never release the backlog.
398 #
399 # repl-backlog-ttl 3600
400
401 # The slave priority is an integer number published by Redis in the INFO output.
402 # It is used by Redis Sentinel in order to select a slave to promote into a
403 # master if the master is no longer working correctly.
404 #
405 # A slave with a low priority number is considered better for promotion, so
406 # for instance if there are three slaves with priority 10, 100, 25 Sentinel will
407 # pick the one with priority 10, that is the lowest.
408 #
409 # However a special priority of 0 marks the slave as not able to perform the
410 # role of master, so a slave with priority of 0 will never be selected by
411 # Redis Sentinel for promotion.
412 #
413 # By default the priority is 100.
414 slave-priority 100
415
416 # It is possible for a master to stop accepting writes if there are less than
417 # N slaves connected, having a lag less or equal than M seconds.
418 #
419 # The N slaves need to be in "online" state.
420 #
421 # The lag in seconds, that must be remove any key according to the LRU algorithm
544 # volatile-random -> remove a random key with an expire set
545 # allkeys-random -> remove a random key, any key
546 # volatile-ttl -> remove the key with the nearest expire time (minor TTL)
547 # noeviction -> don't expire at all, just return an error on write operations
548 #
549 # Note: with any of the above policies, Redis will return an error on write
550 # operations, when there are no suitable keys for eviction.
551 #
552 # At the date of writing these commands are: set setnx setex append
553 # incr decr rpush lpush rpushx lpushx linsert lset rpoplpush sadd
554 # sinter sinterstore sunion sunionstore sdiff sdiffstore zadd zincrby
555 # zunionstore zinterstore hset hsetnx hmset hincrby incrby decrby
556 # getset mset msetnx exec sort
557 #
558 # The default is:
559 #
560 # maxmemory-policy noeviction
561
562 # LRU and minimal TTL algorithms are not precise algorithms but approximated
563 # algorithms (in order to save memory), so you can tune it for speed or
564 # accuracy. For default Redis will check five keys and pick the one that was
565 # used less recently, you can change the sample size using the following
566 # configuration directive.
567 #
568 # The default of 5 produces good enough results. 10 Approximates very closely
569 # true LRU but costs a bit more CPU. 3 is very fast but not very accurate.
570 #
571 # maxmemory-samples 5
572
- maxclients 10000 同一时间内最大clients连接的数量，超过数量的连接会返回一个错误信息
- maxmemory 设置最大内存,如果内存使用量到达了最大内存设置，有6种处理方法：
1、volatile-lru -> remove the key with an expire set using an LRU algorithm（使用LRU算法删除具有过期集的密钥） 2、allkeys-lru -> remove any key according to the LRU algorithm（根据LRU算法删除任何密钥） 3、volatile-random -> remove a random key with an expire set（删除带有过期集的随机密钥） 4、allkeys-random -> remove a random key, any key（删除一个随机键，任意键） 5、volatile-ttl -> remove the key with the nearest expire time (minor TTL) 删除过期时间最近的密钥（次要TTL） 6、noeviction -> don’t expire at all, just return an error on write operations（根本不过期，只需返回一个写操作错误）
默认的设置是 maxmemory-policy noeviction
- maxmemory-samples 5 LRU算法检查的keys个数
573 ############################## APPEND ONLY MODE ###############################
574
575 # By default Redis asynchronously dumps the dataset on disk. This mode is
576 # good enough in many applications, but an issue with the Redis process or
577 # a power outage may result into a few minutes of writes lost (depending on
578 # the configured save points).
579 #
580 # The Append Only File is an alternative persistence mode that provides
581 # much better durability. For instance using the default data fsync policy
582 # (see later in the config file) Redis can lose just one second of writes in a
583 # dramatic event like a server power outage, or a single write if something
584 # wrong with the Redis process itself happens, but the operating system is
585 # still running correctly.
586 #
587 # AOF and RDB persistence can be enabled at the same time without problems.
588 # If the AOF is enabled on startup Redis will load the AOF, that is the file
589 # with the better durability guarantees.
590 #
591 # Please check http://redis.io/topics/persistence for more information.
592
593 appendonly no
594
595 # The name of the append only file (default: "appendonly.aof")
596
597 appendfilename "appendonly.aof"
598
599 # The fsync() call tells the Operating System to actually write data on disk
600 # instead of waiting for more data in the output buffer. Some OS will really flush
601 # data on disk, some other OS will just try to do it ASAP.
602 #
603 # Redis supports three different modes:
604 #
605 # no: don't fsync, just let the OS flush the data when it wants. Faster.
606 # always: fsync after every write to the append only log. Slow, Safest.
607 # everysec: fsync only one time every second. Compromise.
608 #
609 # The default is "everysec", as that's usually the right compromise between
610 # speed and data safety. It's up to you to understand if you can relax this to
611 # "no" that will let the operating system flush the output buffer when
612 # it wants, for better performances (but if you can live with the idea of
613 # some data loss consider the default persistence mode that's snapshotting),
614 # or on the contrary, use "always" that's very slow but a bit safer than
615 # everysec.
616 #
617 # More details please check the following article:
618 # http://antirez.com/post/redis-persistence-demystified.html
619 #
620 # If unsure, use "everysec".
621
622 # appendfsync always
623 appendfsync everysec
624 # appendfsync no
625
626 # When the AOF fsync policy is set to always or everysec, and a background
627 # saving process (a background save or AOF log background rewriting) is
628 # performing a lot of I/O against the disk, in some Linux configurations
629 # Redis may block too long on the fsync() call. Note that there is no fix for
630 # this currently, as even performing fsync in a different thread will block
631 # our synchronous write(2) call.
632 #
633 # In order to mitigate this problem it's possible to use the following option
634 # that will prevent fsync() from being called in the main process while a
635 # BGSAVE or BGREWRITEAOF is in progress.
636 #
637 # This means that while another child is saving, the durability of Redis is
638 # the same as "appendfsync none". In practical terms, this means that it is
639 # possible to lose up to 30 seconds of log in the worst scenario (with the
640 # default Linux settings).
641 #
642 # If you have latency problems turn this to "yes". Otherwise leave it as
643 # "no" that is the safest pick from the point of view of durability.
644
645 no-appendfsync-on-rewrite no
646
647 # Automatic rewrite of the append only file.
648 # Redis is able to automatically rewrite the log file implicitly calling
649 # BGREWRITEAOF when the AOF log size grows by the specified percentage.
650 #
651 # This is how it works: Redis remembers the size of the AOF file after the
652 # latest rewrite (if no rewrite has happened since the restart, the size of
653 # the AOF at startup is used).
654 #
655 # This base size is compared to the current size. If the current size is
656 # bigger than the specified percentage, the rewrite is triggered. Also
657 # you need to specify a minimal size for the AOF file to be rewritten, this
658 # is useful to avoid rewriting the AOF file even if the percentage increase
659 # is reached but it is still pretty small.
660 #
661 # Specify a percentage of zero in order to disable the automatic AOF
662 # rewrite feature.
663
664 auto-aof-rewrite-percentage 100
665 auto-aof-rewrite-min-size 64mb
666
667 # An AOF file may be found to be truncated at the end during the Redis
668 # startup process, when the AOF data gets loaded back into memory.
669 # This may happen when the system where Redis is running
670 # crashes, especially when an ext4 filesystem is mounted without the
671 # data=ordered option (however this can't happen when Redis itself
672 # crashes or aborts but the operating system still works correctly).
673 #
674 # Redis can either exit with an error when this happens, or load as much
675 # data as possible (the default now) and start if the AOF file is found
676 # to be truncated at the end. The following option controls this behavior.
677 #
678 # If aof-load-truncated is set to yes, a truncated AOF file is loaded and
679 # the Redis server starts emitting a log to inform the user of the event.
680 # Otherwise if the option is set to no, the server aborts with an error
681 # and refuses to start. When the option is set to no, the user requires
682 # to fix the AOF file using the "redis-check-aof" utility before to restart
683 # the server.
684 #
685 # Note that if the AOF file will be found to be corrupted in the middle
686 # the server will still exit with an error. This option only applies when
687 # Redis will try to read more data from the AOF file but not enough bytes
688 # will be found.
689 aof-load-truncated yes
690
- appendonly yes 启用AOF模式
- appendfilename “appendonly.aof” 设置AOF记录的文件名
- appendfsync always 同步持久化 每次发生数据变更会被立即记录到磁盘 性能较差但数据完整性比较好
- appendfsync everysec 出厂默认推荐，异步操作，每秒记录 如果一秒内宕机，有数据丢失
- appendfsync no 向磁盘进行数据刷写的频率,按照OS自身的刷写策略来进行，速度最快
- no-appendfsync-on-rewrite no 当主进程在进行向磁盘的写操作时，将会阻止其它的fsync调用
- auto-aof-rewrite-percentage 100 aof文件触发自动rewrite的百分比，值为0则表示禁用自动rewrite
- auto-aof-rewrite-min-size 64mb aof文件触发自动rewrite的最小文件size
- aof-load-truncated yes 是否加载不完整的aof文件来进行启动
691 ################################ LUA SCRIPTING ###############################
692
693 # Max execution time of a Lua script in milliseconds.
694 #
695 # If the maximum execution time is reached Redis will log that a script is
696 # still in execution after the maximum allowed time and will start to
697 # reply to queries with an error.
698 #
699 # When a long running script exceeds the maximum execution time only the
700 # SCRIPT KILL and SHUTDOWN NOSAVE commands are available. The first can be
701 # used to stop a script that did not yet called write commands. The second
702 # is the only way to shut down the server in the case a write command was
703 # already issued by the script but the user doesn't want to wait for the natural
704 # termination of the script.
705 #
706 # Set it to 0 or a negative value for unlimited execution without warnings.
707 lua-time-limit 5000
708
- lua-time-limit 5000 设置lua脚本的最大运行时间，单位为毫秒
709 ################################ REDIS CLUSTER ###############################
710 #
711 # ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
712 # WARNING EXPERIMENTAL: Redis Cluster is considered to be stable code, however
713 # in order to mark it as "mature" we need to wait for a non trivial percentage
714 # of users to deploy it in production.
715 # ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
716 #
717 # Normal Redis instances can't be part of a Redis Cluster; only nodes that are
718 # started as cluster nodes can. In order to start a Redis instance as a
719 # cluster node enable the cluster support uncommenting the following:
720 #
721 # cluster-enabled yes
722
723 # Every cluster node has a cluster configuration file. This file is not
724 # intended to be edited by hand. It is created and updated by Redis nodes.
725 # Every Redis Cluster node requires a different cluster configuration file.
726 # Make sure that instances running in the same system do not have
727 # overlapping cluster configuration file names.
728 #
729 # cluster-config-file nodes-6379.conf
730
731 # Cluster node timeout is the amount of milliseconds a node must be unreachable
732 # for it to be considered in failure state.
733 # Most other internal time limits are multiple of the node timeout.
734 #
735 # cluster-node-timeout 15000
736
737 # A slave of a failing master will avoid to start a failover if its data
738 # looks too old.
739 #
740 # There is no simple way for a slave to actually have a exact measure of
741 # its "data age", so the following two checks are performed:
742 #
743 # 1) If there are multiple slaves able to failover, they exchange messages
744 # in order to try to give an advantage to the slave with the best
745 # replication offset (more data from the master processed).
746 # Slaves will try to get their rank by offset, and apply to the start
747 # of the failover a delay proportional to their rank.
748 #
749 # 2) Every single slave computes the time of the last interaction with
750 # its master. This can be the last ping or command received (if the master
751 # is still in the "connected" state), or the time that elapsed since the
752 # disconnection with the master (if the replication link is currently down).
753 # If the last interaction is too old, the slave will not try to failover
754 # at all.
755 #
756 # The point "2" can be tuned by user. Specifically a slave will not perform
757 # the failover if, since the last interaction with the master, the time
758 # elapsed is greater than:
759 #
760 # (node-timeout * slave-validity-factor) + repl-ping-slave-period
761 #
762 # So for example if node-timeout is 30 seconds, and the slave-validity-factor
763 # is 10, and assuming a default repl-ping-slave-period of 10 seconds, the
764 # slave will not try to failover if it was not able to talk with the master
765 # for longer than 310 seconds.
766 #
767 # A large slave-validity-factor may allow slaves with too old data to failover
768 # a master, while a too small value may prevent the cluster from being able to
769 # elect a slave at all.
770 #
771 # For maximum availability, it is possible to set the slave-validity-factor
772 # to a value of 0, which means, that slaves will always try to failover the
773 # master regardless of the last time they interacted with the master.
774 # (However they'll always try to apply a delay proportional to their
775 # offset rank).
776 #
777 # Zero is the only value able to guarantee that when all the partitions heal
778 # the cluster will always be able to continue.
779 #
780 # cluster-slave-validity-factor 10
781
782 # Cluster slaves are able to migrate to orphaned masters, that are masters
783 # that are left without working slaves. This improves the cluster ability
784 # to resist to failures as otherwise an orphaned master can't be failed over
785 # in case of failure if it has no working slaves.
786 #
787 # Slaves migrate to orphaned masters only if there are still at least a
788 # given number of other working slaves for their old master. This number
789 # is the "migration barrier". A migration barrier of 1 means that a slave
790 # will migrate only if there is at least 1 other working slave for its master
791 # and so forth. It usually reflects the number of slaves you want for every
792 # master in your cluster.
793 #
794 # Default is 1 (slaves migrate only if their masters remain with at least
795 # one slave). To disable migration just set it to a very large value.
796 # A value of 0 can be set but is useful only for debugging and dangerous
797 # in production.
798 #
799 # cluster-migration-barrier 1
800
801 # By default Redis Cluster nodes stop accepting queries if they detect there
802 # is at least an hash slot uncovered (no available node is serving it).
803 # This way if the cluster is partially down (for example a range of hash slots
804 # are no longer covered) all the cluster becomes, eventually, unavailable.
805 # It automatically returns available as soon as all the slots are covered again.
806 #
807 # However sometimes you want the subset of the cluster which is working,
808 # to continue to accept queries for the part of the key space that is still
809 # covered. In order to do so, just set the cluster-require-full-coverage
810 # option to no.
811 #
812 # cluster-require-full-coverage yes
813
814 # In order to setup your cluster make sure to read the documentation
815 # available at http://redis.io web site.
816
- cluster-enabled yes 配置redis做为一个集群节点来启动
- cluster-config-file node-6379.conf 每个集群节点都有一个集群配置文件，这个文件不需要编辑，它由redis节点来创建和更新。每个redis节点的集群配置文件不可以相同。
- cluster-node-timeout 15000 设置集群节点超时时间，如果超过了指定的超时时间后仍不可达，则节点被认为是失败状态，单位为毫秒。
- （node-timeout * slave-validity-factor）+repl-ping-slave-period 一个比较大的slave-validity-factor参数能够允许slave端使用比较旧的数据去failover它的master端，而一个比较小的值可能会阻止集群去选择slave端。为获得最大的可用性，可以设置slave-validity-factor的值为0，这表示slave端将会一直去尝试failover它的master端而不管它与master端的最后交互时间。
- cluster-slave-validity-factor 10 默认值为10,集群中的slave可以迁移到那些没有可用slave的master端，这提升了集群处理故障的能力。毕竟一个没有slave的master端如果发生了故障是没有办法去进行failover的。要将一个slave迁移到别的master，必须这个slave的原master端有至少给定数目的可用slave才可以进行迁移，这个给定的数目由migration barrier参数来进行设置，默认值为1，表示这个要进行迁移的slave的原master端应该至少还有1个可用的slave才允许其进行迁移，要禁用这个功能只需要将此参数设置为一个非常大的值。
- cluster-migration-barrier 1 默认情况下当redis集群节点发现有至少一个hashslot未被covered时将会停止接收查询。这种情况下如果有一部份的集群down掉了，那整个集群将变得不可用。集群将会在所有的slot重新covered之后自动恢复可用。若想要设置集群在部份key space没有cover完成时继续去接收查询，就将参数设置为no。
- cluster-require-full-coverage yes
817 ################################## SLOW LOG ###################################
818
819 # The Redis Slow Log is a system to log queries that exceeded a specified
820 # execution time. The execution time does not include the I/O operations
821 # like talking with the client, sending the reply and so forth,
822 # but just the time needed to actually execute the command (this is the only
823 # stage of command execution where the thread is blocked and can not serve
824 # other requests in the meantime).
825 #
826 # You can configure the slow log with two parameters: one tells Redis
827 # what is the execution time, in microseconds, to exceed in order for the
828 # command to get logged, and the other parameter is the length of the
829 # slow log. When a new command is logged the oldest one is removed from the
830 # queue of logged commands.
831
832 # The following time is expressed in microseconds, so 1000000 is equivalent
833 # to one second. Note that a negative number disables the slow log, while
834 # a value of zero forces the logging of every command.
835 slowlog-log-slower-than 10000
836
837 # There is no limit to this length. Just be aware that it will consume memory.
838 # You can reclaim memory used by the slow log with SLOWLOG RESET.
839 slowlog-max-len 128
840
redis的slow log是一个系统OS进行的记录查询，它是超过了指定的执行时间的。执行时间不包括类似与client进行交互或发送回复等I/O操作，它只是实际执行指令的时间。
- slowlog-log-slower-than 10000 告诉redis执行时间，这个时间是微秒级的（1秒=1000000微秒），这是为了不遗漏命令
- slowlog-max-len 128 设置slowlog的长度，当一个新的命令被记录时，最旧的命令将会从命令记录队列中移除。
841 ################################ LATENCY MONITOR ##############################
842
843 # The Redis latency monitoring subsystem samples different operations
844 # at runtime in order to collect data related to possible sources of
845 # latency of a Redis instance.
846 #
847 # Via the LATENCY command this information is available to the user that can
848 # print graphs and obtain reports.
849 #
850 # The system only logs operations that were performed in a time equal or
851 # greater than the amount of milliseconds specified via the
852 # latency-monitor-threshold configuration directive. When its value is set
853 # to zero, the latency monitor is turned off.
854 #
855 # By default latency monitoring is disabled since it is mostly not needed
856 # if you don't have latency issues, and collecting data has a performance
857 # impact, that while very small, can be measured under big load. Latency
858 # monitoring can easily be enabled at runtime using the command
859 # "CONFIG SET latency-monitor-threshold " if needed.
860 latency-monitor-threshold 0
861
- latency-monitor-threshold 0 延迟监控，用于记录等于或超过了指定时间的操作，默认是关闭状态，即值为0。
862 ############################# EVENT NOTIFICATION ##############################
863
864 # Redis can notify Pub/Sub clients about events happening in the key space.
865 # This feature is documented at http://redis.io/topics/notifications
866 #
867 # For instance if keyspace events notification is enabled, and a client
868 # performs a DEL operation on key "foo" stored in the Database 0, two
869 # messages will be published via Pub/Sub:
870 #
871 # PUBLISH __keyspace@0__:foo del
872 # PUBLISH __keyevent@0__:del foo
873 #
874 # It is possible to select the events that Redis will notify among a set
875 # of classes. Every class is identified by a single character:
876 #
877 # K Keyspace events, published with __keyspace@__ prefix.
878 # E Keyevent events, published with __keyevent@__ prefix.
879 # g Generic commands (non-type specific) like DEL, EXPIRE, RENAME, ...
880 # $ String commands
881 # l List commands
882 # s Set commands
883 # h Hash commands
884 # z Sorted set commands
885 # x Expired events (events generated every time a key expires)
886 # e Evicted events (events generated when a key is evicted for maxmemory)
887 # A Alias for g$lshzxe, so that the "AKE" string means all the events.
888 #
889 # The "notify-keyspace-events" takes as argument a string that is composed
890 # of zero or multiple characters. The empty string means that notifications
891 # are disabled.
892 #
893 # Example: to enable list and generic events, from the point of view of the
894 # event name, use:
895 #
896 # notify-keyspace-events Elg
897 #
898 # Example 2: to get the stream of the expired keys subscribing to channel
899 # name __keyevent@0__:expired use:
900 #
901 # notify-keyspace-events Ex
902 #
903 # By default all notifications are disabled because most users don't need
904 # this feature and the feature has some overhead. Note that if you don't
905 # specify at least one of K or E, no events will be delivered.
906 notify-keyspace-events ""
907
- notify-keyspace-events “” 事件通知，默认不启用，具体参数查看配置文件
908 ############################### ADVANCED CONFIG ###############################
909
910 # Hashes are encoded using a memory efficient data structure when they have a
911 # small number of entries, and the biggest entry does not exceed a given
912 # threshold. These thresholds can be configured using the following directives.
913 hash-max-ziplist-entries 512
914 hash-max-ziplist-value 64
915
916 # Lists are also encoded in a special way to save a lot of space.
917 # The number of entries allowed per internal list node can be specified
918 # as a fixed maximum size or a maximum number of elements.
919 # For a fixed maximum size, use -5 through -1, meaning:
920 # -5: max size: 64 Kb node->node->...->node->[tail]
939 # [head], [tail] will always be uncompressed; inner nodes will compress.
940 # 2: [head]->[next]->node->node->...->node->[prev]->[tail]
941 # 2 here means: don't compress head or head->next or tail->prev or tail,
942 # but compress all nodes between them.
943 # 3: [head]->[next]->[next]->node->node->...->node->[prev]->[prev]->[tail]
944 # etc.
945 list-compress-depth 0
946
947 # Sets have a special encoding in just one case: when a set is composed
948 # of just strings that happen to be integers in radix 10 in the range
949 # of 64 bit signed integers.
950 # The following configuration setting sets the limit in the size of the
951 # set in order to use this special memory saving encoding.
952 set-max-intset-entries 512
953
954 # Similarly to hashes and lists, sorted sets are also specially encoded in
955 # order to save a lot of space. This encoding is only used when the length and
956 # elements of a sorted set are below the following limits:
957 zset-max-ziplist-entries 128
958 zset-max-ziplist-value 64
959
960 # HyperLogLog sparse representation bytes limit. The limit includes the
961 # 16 bytes header. When an HyperLogLog using the sparse representation crosses
962 # this limit, it is converted into the dense representation.
963 #
964 # A value greater than 16000 is totally useless, since at that point the
965 # dense representation is more memory efficient.
966 #
967 # The suggested value is ~ 3000 in order to have the benefits of
968 # the space efficient encoding without slowing down too much PFADD,
969 # which is O(N) with the sparse encoding. The value can be raised to
970 # ~ 10000 when CPU is not a concern, but space is, and the data set is
971 # composed of many HyperLogLogs with cardinality in the 0 - 15000 range.
972 hll-sparse-max-bytes 3000
973
974 # Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in
975 # order to help rehashing the main Redis hash table (the one mapping top-level
976 # keys to values). The hash table implementation Redis uses (see dict.c)
977 # performs a lazy rehashing: the more operation you run into a hash table
978 # that is rehashing, the more rehashing "steps" are performed, so if the
979 # server is idle the rehashing is never complete and some more memory is used
980 # by the hash table.
981 #
982 # The default is to use this millisecond 10 times every second in order to
983 # actively rehash the main dictionaries, freeing memory when possible.
984 #
985 # If unsure:
986 # use "activerehashing no" if you have hard latency requirements and it is
987 # not a good thing in your environment that Redis can reply from time to time
988 # to queries with 2 milliseconds delay.
989 #
990 # use "activerehashing yes" if you don't have such hard requirements but
991 # want to free memory asap when possible.
992 activerehashing yes
993
994 # The client output buffer limits can be used to force disconnection of clients
995 # that are not reading data from the server fast enough for some reason (a
996 # common reason is that a Pub/Sub client can't consume messages as fast as the
997 # publisher can produce them).
998 #
999 # The limit can be set differently for the three different classes of clients:
1000 #
1001 # normal -> normal clients including MONITOR clients
1002 # slave -> slave clients
1003 # pubsub -> clients subscribed to at least one pubsub channel or pattern
1004 #
1005 # The syntax of every client-output-buffer-limit directive is the following:
1006 #
1007 # client-output-buffer-limit
1008 #
1009 # A client is immediately disconnected once the hard limit is reached, or if
1010 # the soft limit is reached and remains reached for the specified number of
1011 # seconds (continuously).
1012 # So for instance if the hard limit is 32 megabytes and the soft limit is
1013 # 16 megabytes / 10 seconds, the client will get disconnected immediately
1014 # if the size of the output buffers reach 32 megabytes, but will also get
1015 # disconnected if the client reaches 16 megabytes and continuously overcomes
1016 # the limit for 10 seconds.
1017 #
1018 # By default normal clients are not limited because they don't receive data
1019 # without asking (in a push way), but just after a request, so only
1020 # asynchronous clients may create a scenario where data is requested faster
1021 # than it can read.
1022 #
1023 # Instead there is a default limit for pubsub and slave clients, since
1024 # subscribers and slaves receive data in a push fashion.
1025 #
1026 # Both the hard or the soft limit can be disabled by setting them to zero.
1027 client-output-buffer-limit normal 0 0 0
1028 client-output-buffer-limit slave 256mb 64mb 60
1029 client-output-buffer-limit pubsub 32mb 8mb 60
1030
1031 # Redis calls an internal function to perform many background tasks, like
1032 # closing connections of clients in timeout, purging expired keys that are
1033 # never requested, and so forth.
1034 #
1035 # Not all tasks are performed with the same frequency, but Redis checks for
1036 # tasks to perform according to the specified "hz" value.
1037 #
1038 # By default "hz" is set to 10. Raising the value will use more CPU when
1039 # Redis is idle, but at the same time will make Redis more responsive when
1040 # there are many keys expiring at the same time, and timeouts may be
1041 # handled with more precision.
1042 #
1043 # The range is between 1 and 500, however a value over 100 is usually not
1044 # a good idea. Most users should use the default of 10 and raise this up to
1045 # 100 only in environments where very low latency is required.
1046 hz 10
1047
1048 # When a child rewrites the AOF file, if the following option is enabled
1049 # the file will be fsync-ed every 32 MB of data generated. This is useful
1050 # in order to commit the file to the disk more incrementally and avoid
1051 # big latency spikes.
1052 aof-rewrite-incremental-fsync yes
-
当条目数量较少且最大不会超过给定阀值时，哈希编码将使用一个很高效的内存数据结构，阀值由以下参数来进行配置。 hash-max-ziplist-entries 512 hash-max-ziplist-value 64
-
与哈希类似，少量的lists也会通过一个指定的方式去编码从而节省更多的空间，它的阀值通过以下参数来进行配置。 list-max-ziplist-entries 512 list-max-ziplist-value 64
-
集合sets在一种特殊的情况时有指定的编码方式，这种情况是集合由一组10进制的64位有符号整数范围内的数字组成的情况。以下选项可以设置集合使用这种特殊编码方式的size限制。 set-max-intset-entries 512
-
与哈希和列表类似，有序集合也会使用一种特殊的编码方式来节省空间，这种特殊的编码方式只用于这个有序集合的长度和元素均低于以下参数设置的值时。 zset-max-ziplist-entries 128 zset-max-ziplist-value 64
-
hll-sparse-max-bytes 3000 设置HyeperLogLog的字节数限制，这个值通常在0~15000之间，默认为3000，基本不超过16000
-
activerehashing yes redis将会在每秒中抽出10毫秒来对主字典进行重新散列化处理，这有助于尽可能的释放内存 因为某些原因，client不能足够快的从server读取数据，那client的输出缓存限制可能会使client失连，这个限制可用于3种不同的client种类，分别是：normal、slave和pubsub。 进行设置的格式如下：
client-output-buffer-limit
如果达到hard limit那client将会立即失连。 如果达到soft limit那client将会在soft seconds秒之后失连。 参数soft limit < hard limit。
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使用一个内部程序来处理后台任务，例如关闭超时的client连接，清除过期的key等等。它并不会同时处理所有的任务，redis通过指定的hz参数去检查和执行任务。 hz默认设为10，提高它的值将会占用更多的cpu，当然相应的redis将会更快的处理同时到期的许多key，以及更精确的去处理超时。 hz的取值范围是1~500，通常不建议超过100，只有在请求延时非常低的情况下可以将值提升到100。 hz 10 当一个子进程要改写AOF文件，如果以下选项启用，那文件将会在每产生32MB数据时进行同步，这样提交增量文件到磁盘时可以避免出现比较大的延迟。 aof-rewrite-incremental-fsync yes
参考： https://www.cnblogs.com/kingsonfu/p/10138647.html https://www.cnblogs.com/zxtceq/p/7676911.html https://www.cnblogs.com/ysocean/p/9074787.html http://blog.chinaunix.net/uid-20682147-id-5816952.html https://blog.csdn.net/yfkiss/article/details/21476801 https://blog.csdn.net/ljl890705/article/details/51540427
