使用AOF持久化文件实现还原Redis数据库并得到RDB持久化文件
目录
1 编写本文的初衷
因为目前实习工作需求,需要把服务器环境中所有Redis数据进行初步简单分析,即统计其中存储的每一个key所占内存的大小,以便作出清理不重要缓存数据的决策。
但是,由于从线上环境获得持久化文件为AOF文件,而不是RDB文件。RDB文件可以通过Rdbtools工具,来分析具体数据。但是AOF文件不能这样操作。
因此,就给我带来一个问题:如何通过AOF文件获取指定的RDB持久化文件呢?
于是,我通过查阅网上文章,获取的一个解决思路:单独在Redis中开启一个未使用过的端口服务,使用已得到的AOF文件替换该端口服务下自动生成的AOF文件;然后,重启该端口指定的Redis服务,即可把新的AOF文件中数据加载到Redis数据库中,最后在该端口服务客户端执行save或者bgsave命令,即可在指定路径下得到对应的RDB持久化文件。
2 具体实施
2.1 Redis持久化概念简介
Redis数据库进行持久化有两种方式:RDB持久化和AOF持久化。
那么,什么是RDB持久化呢?
RDB(Redis Database)持久化:可以将Redis在内存中的数据库状态保存到磁盘里面,避免数据意外丢失。RDB持久化既可以手动执行,也可以根据服务器配置选项定期执行,该功能可以将某个时间点上的数据库状态保存到一个RDB文件中。(PS:手动执行保存时,在客户端执行SAVE命令或者BGSAVE即可把当前所有数据保存到dump.rdb文件中,如果在线上执行,建议使用BGSAVE命令)
RDB文件具体功能:用于保存和还原Redis服务器所有数据库中的所有键值对数据。
那么,什么是AOF持久化呢?
AOF(Append Only File)持久化:与RDB持久化通过保存数据库中的键值对来记录数据库状态不同,AOF持久化是通过保存Redis服务器所执行的写命令来记录数据库状态的。AOF持久化功能的实现可以分为命令追加(append)、文件写入、文件同步(sync)三个步骤。
AOF文件具体功能:通过保存所有修改数据库的写命令请求来记录服务器的数据库状态。
2.2 获取指定Redis的AOF持久化文件
一般情况,都是获取限制环境的AOF文件,那么如何在线上环境找到AOF文件呢?(PS:因为时间原因,可能忘记存储在哪里,所以以下提供一个搜索命令,方便操作)
- sudo find / -name '*.aof' # 此命令用于查找系统上所有以aof为后缀的文件
通过该命令,查看具体文件的路径信息,即可确认自己需要获取的AOF文件。
确定后,通过一下命令把指定AOF文件拷贝到本地主机上:
- scp facelive@172.169.18.14:/home/facelive/prod_redis_data/redis/db-appendonly.aof . # 从服务器复制远程文件到本地当前所在根目录
2.3 把Redis的持久化AOF文件转换为RDB文件
关于redis.conf配置aof持久化文件读取修改配置简单介绍:
(1)找到redis.conf文件,设置其中的字段属性:
- appendonly no ——> appendonly yes
此处也可以在redis客户端,使用指令来完成修改:
- redis 127.0.0.1:6379> config set appendonly yes
- OK
- redis 127.0.0.1:6379> BGREWRITEAOF # 用于重写生成aof文件
- Background append only file rewriting started
此选项为aof功能的开关,默认为“no”,可以通过“yes”来开启aof功能
只有在“yes”下,aof重写/文件同步等特性才会生效
(2)在redis.conf文件中,指定aof文件的名称
- appendfilename "appendonly.aof" # 这是文件中默认的配置名称,也可以自己修改指定的文件名称
(3)在redis.conf文件中,确认 aof操作中文件同步策略
配置默认结果:
- # appendfsync always
- appendfsync everysec
- # appendfsync no
即选用everysec,具体意思:
1. no:表示等操作系统进行数据缓存同步到磁盘.
2. always:表示每次更新操作后手动调用fsync() 将数据写到磁盘.
3. everysec:表示每秒同步一次.一般用everysec
(4)在redis.conf文件中,确认 aof-rewrite期间,appendfsync是否暂缓文件同步
配置默认结果:
- no-appendfsync-on-rewrite no
具体意思:
“no”表示“不暂缓”,“yes”表示“暂缓”,默认为“no”
(5)在redis.conf文件中,确认 aof文件rewrite触发的最小文件尺寸(mb,gb),以及 相对于“上一次”rewrite,本次rewrite触发时aof文件应该增长的百分比
配置默认结果:
- auto-aof-rewrite-percentage 100
- auto-aof-rewrite-min-size 64mb
具体实施步骤:
(1) 创建一个新的redis.conf文件,该文件命名可采用redis_port.conf形式,例如:redis_6391.conf。该文件中内容起初完全何Redis默认的redis.conf文件中内容一致
(2) 修改redis_6391.conf指定的port值,在文件中搜索port把默认的6379修改为6391
(3) 修改redis_6391指定的dir值,在文件中搜索dir把默认的”.\”改为自己要存放文件的具体路径。该路径用于存放RDB文件和AOF文件
(4) 修改redis_6391指定的appendfilename值,在文件中搜索appendfilename把默认的”appendonly.aof”改为自己想要定义的文件名称,该文件即为AOF文件的最终名称
(5) 修改redis_6391指定的dbfilename值,在文件中搜索dbfilename把默认的”dump.rdb”改为自己想要定义的文件名称,该文件即为RDB文件的最终名称
(6) 此步骤最重要,修改redis_6391指定的appendonly值,在文件中搜索appendonly把默认的”no”改为”yes”。这句配置意思是Redis服务重启后,默认不加载AOF持久化文件恢复数据,而是去找RDB持久化文件恢复;如果修改为”yes”后,发现有AOF文件,会首先加载AOF文件恢复数据
以下给出我本机修改后的redis_6391.conf文件中具体配置代码:
- 1 # Redis configuration file example
- 2
- 3 # Note on units: when memory size is needed, it is possible to specify
- 4 # it in the usual form of 1k 5GB 4M and so forth:
- 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 # units are case insensitive so 1GB 1Gb 1gB are all the same.
- 14
- 15 ################################## INCLUDES ###################################
- 16
- 17 # Include one or more other config files here. This is useful if you
- 18 # have a standard template that goes to all Redis server but also need
- 19 # to customize a few per-server settings. Include files can include
- 20 # other files, so use this wisely.
- 21 #
- 22 # Notice option "include" won't be rewritten by command "CONFIG REWRITE"
- 23 # from admin or Redis Sentinel. Since Redis always uses the last processed
- 24 # line as value of a configuration directive, you'd better put includes
- 25 # at the beginning of this file to avoid overwriting config change at runtime.
- 26 #
- 27 # If instead you are interested in using includes to override configuration
- 28 # options, it is better to use include as the last line.
- 29 #
- 30 # include /path/to/local.conf
- 31 # include /path/to/other.conf
- 32
- 33 ################################ GENERAL #####################################
- 34
- 35 # By default Redis does not run as a daemon. Use 'yes' if you need it.
- 36 # Note that Redis will write a pid file in /var/run/redis.pid when daemonized.
- 37 daemonize no
- 38
- 39 # When running daemonized, Redis writes a pid file in /var/run/redis.pid by
- 40 # default. You can specify a custom pid file location here.
- 41 pidfile /var/run/redis.pid
- 42
- 43 # Accept connections on the specified port, default is 6391.
- 44 # If port 0 is specified Redis will not listen on a TCP socket.
- 45 port 6391
- 46
- 47 # TCP listen() backlog.
- 48 #
- 49 # In high requests-per-second environments you need an high backlog in order
- 50 # to avoid slow clients connections issues. Note that the Linux kernel
- 51 # will silently truncate it to the value of /proc/sys/net/core/somaxconn so
- 52 # make sure to raise both the value of somaxconn and tcp_max_syn_backlog
- 53 # in order to get the desired effect.
- 54 tcp-backlog 511
- 55
- 56 # By default Redis listens for connections from all the network interfaces
- 57 # available on the server. It is possible to listen to just one or multiple
- 58 # interfaces using the "bind" configuration directive, followed by one or
- 59 # more IP addresses.
- 60 #
- 61 # Examples:
- 62 #
- 63 # bind 192.168.1.100 10.0.0.1
- 64 # bind 127.0.0.1
- 65
- 66 # Specify the path for the Unix socket that will be used to listen for
- 67 # incoming connections. There is no default, so Redis will not listen
- 68 # on a unix socket when not specified.
- 69 #
- 70 # unixsocket /tmp/redis.sock
- 71 # unixsocketperm 700
- 72
- 73 # Close the connection after a client is idle for N seconds (0 to disable)
- 74 timeout 0
- 75
- 76 # TCP keepalive.
- 77 #
- 78 # If non-zero, use SO_KEEPALIVE to send TCP ACKs to clients in absence
- 79 # of communication. This is useful for two reasons:
- 80 #
- 81 # 1) Detect dead peers.
- 82 # 2) Take the connection alive from the point of view of network
- 83 # equipment in the middle.
- 84 #
- 85 # On Linux, the specified value (in seconds) is the period used to send ACKs.
- 86 # Note that to close the connection the double of the time is needed.
- 87 # On other kernels the period depends on the kernel configuration.
- 88 #
- 89 # A reasonable value for this option is 60 seconds.
- 90 tcp-keepalive 0
- 91
- 92 # Specify the server verbosity level.
- 93 # This can be one of:
- 94 # debug (a lot of information, useful for development/testing)
- 95 # verbose (many rarely useful info, but not a mess like the debug level)
- 96 # notice (moderately verbose, what you want in production probably)
- 97 # warning (only very important / critical messages are logged)
- 98 loglevel notice
- 99
- 100 # Specify the log file name. Also the empty string can be used to force
- 101 # Redis to log on the standard output. Note that if you use standard
- 102 # output for logging but daemonize, logs will be sent to /dev/null
- 103 logfile ""
- 104
- 105 # To enable logging to the system logger, just set 'syslog-enabled' to yes,
- 106 # and optionally update the other syslog parameters to suit your needs.
- 107 # syslog-enabled no
- 108
- 109 # Specify the syslog identity.
- 110 # syslog-ident redis
- 111
- 112 # Specify the syslog facility. Must be USER or between LOCAL0-LOCAL7.
- 113 # syslog-facility local0
- 114
- 115 # Set the number of databases. The default database is DB 0, you can select
- 116 # a different one on a per-connection basis using SELECT <dbid> where
- 117 # dbid is a number between 0 and 'databases'-1
- 118 databases 16
- 119
- 120 ################################ SNAPSHOTTING ################################
- 121 #
- 122 # Save the DB on disk:
- 123 #
- 124 # save <seconds> <changes>
- 125 #
- 126 # Will save the DB if both the given number of seconds and the given
- 127 # number of write operations against the DB occurred.
- 128 #
- 129 # In the example below the behaviour will be to save:
- 130 # after 900 sec (15 min) if at least 1 key changed
- 131 # after 300 sec (5 min) if at least 10 keys changed
- 132 # after 60 sec if at least 10000 keys changed
- 133 #
- 134 # Note: you can disable saving at all commenting all the "save" lines.
- 135 #
- 136 # It is also possible to remove all the previously configured save
- 137 # points by adding a save directive with a single empty string argument
- 138 # like in the following example:
- 139 #
- 140 # save ""
- 141
- 142 save 900 1
- 143 save 300 10
- 144 save 60 10000
- 145
- 146 # By default Redis will stop accepting writes if RDB snapshots are enabled
- 147 # (at least one save point) and the latest background save failed.
- 148 # This will make the user aware (in a hard way) that data is not persisting
- 149 # on disk properly, otherwise chances are that no one will notice and some
- 150 # disaster will happen.
- 151 #
- 152 # If the background saving process will start working again Redis will
- 153 # automatically allow writes again.
- 154 #
- 155 # However if you have setup your proper monitoring of the Redis server
- 156 # and persistence, you may want to disable this feature so that Redis will
- 157 # continue to work as usual even if there are problems with disk,
- 158 # permissions, and so forth.
- 159 stop-writes-on-bgsave-error yes
- 160
- 161 # Compress string objects using LZF when dump .rdb databases?
- 162 # For default that's set to 'yes' as it's almost always a win.
- 163 # If you want to save some CPU in the saving child set it to 'no' but
- 164 # the dataset will likely be bigger if you have compressible values or keys.
- 165 rdbcompression yes
- 166
- 167 # Since version 5 of RDB a CRC64 checksum is placed at the end of the file.
- 168 # This makes the format more resistant to corruption but there is a performance
- 169 # hit to pay (around 10%) when saving and loading RDB files, so you can disable it
- 170 # for maximum performances.
- 171 #
- 172 # RDB files created with checksum disabled have a checksum of zero that will
- 173 # tell the loading code to skip the check.
- 174 rdbchecksum yes
- 175
- 176 # The filename where to dump the DB
- 177 dbfilename dump_6391.rdb
- 178
- 179 # The working directory.
- 180 #
- 181 # The DB will be written inside this directory, with the filename specified
- 182 # above using the 'dbfilename' configuration directive.
- 183 #
- 184 # The Append Only File will also be created inside this directory.
- 185 #
- 186 # Note that you must specify a directory here, not a file name.
- 187 dir /home/liuzhen/data
- 188
- 189 ################################# REPLICATION #################################
- 190
- 191 # Master-Slave replication. Use slaveof to make a Redis instance a copy of
- 192 # another Redis server. A few things to understand ASAP about Redis replication.
- 193 #
- 194 # 1) Redis replication is asynchronous, but you can configure a master to
- 195 # stop accepting writes if it appears to be not connected with at least
- 196 # a given number of slaves.
- 197 # 2) Redis slaves are able to perform a partial resynchronization with the
- 198 # master if the replication link is lost for a relatively small amount of
- 199 # time. You may want to configure the replication backlog size (see the next
- 200 # sections of this file) with a sensible value depending on your needs.
- 201 # 3) Replication is automatic and does not need user intervention. After a
- 202 # network partition slaves automatically try to reconnect to masters
- 203 # and resynchronize with them.
- 204 #
- 205 # slaveof <masterip> <masterport>
- 206
- 207 # If the master is password protected (using the "requirepass" configuration
- 208 # directive below) it is possible to tell the slave to authenticate before
- 209 # starting the replication synchronization process, otherwise the master will
- 210 # refuse the slave request.
- 211 #
- 212 # masterauth <master-password>
- 213
- 214 # When a slave loses its connection with the master, or when the replication
- 215 # is still in progress, the slave can act in two different ways:
- 216 #
- 217 # 1) if slave-serve-stale-data is set to 'yes' (the default) the slave will
- 218 # still reply to client requests, possibly with out of date data, or the
- 219 # data set may just be empty if this is the first synchronization.
- 220 #
- 221 # 2) if slave-serve-stale-data is set to 'no' the slave will reply with
- 222 # an error "SYNC with master in progress" to all the kind of commands
- 223 # but to INFO and SLAVEOF.
- 224 #
- 225 slave-serve-stale-data yes
- 226
- 227 # You can configure a slave instance to accept writes or not. Writing against
- 228 # a slave instance may be useful to store some ephemeral data (because data
- 229 # written on a slave will be easily deleted after resync with the master) but
- 230 # may also cause problems if clients are writing to it because of a
- 231 # misconfiguration.
- 232 #
- 233 # Since Redis 2.6 by default slaves are read-only.
- 234 #
- 235 # Note: read only slaves are not designed to be exposed to untrusted clients
- 236 # on the internet. It's just a protection layer against misuse of the instance.
- 237 # Still a read only slave exports by default all the administrative commands
- 238 # such as CONFIG, DEBUG, and so forth. To a limited extent you can improve
- 239 # security of read only slaves using 'rename-command' to shadow all the
- 240 # administrative / dangerous commands.
- 241 slave-read-only yes
- 242
- 243 # Slaves send PINGs to server in a predefined interval. It's possible to change
- 244 # this interval with the repl_ping_slave_period option. The default value is 10
- 245 # seconds.
- 246 #
- 247 # repl-ping-slave-period 10
- 248
- 249 # The following option sets the replication timeout for:
- 250 #
- 251 # 1) Bulk transfer I/O during SYNC, from the point of view of slave.
- 252 # 2) Master timeout from the point of view of slaves (data, pings).
- 253 # 3) Slave timeout from the point of view of masters (REPLCONF ACK pings).
- 254 #
- 255 # It is important to make sure that this value is greater than the value
- 256 # specified for repl-ping-slave-period otherwise a timeout will be detected
- 257 # every time there is low traffic between the master and the slave.
- 258 #
- 259 # repl-timeout 60
- 260
- 261 # Disable TCP_NODELAY on the slave socket after SYNC?
- 262 #
- 263 # If you select "yes" Redis will use a smaller number of TCP packets and
- 264 # less bandwidth to send data to slaves. But this can add a delay for
- 265 # the data to appear on the slave side, up to 40 milliseconds with
- 266 # Linux kernels using a default configuration.
- 267 #
- 268 # If you select "no" the delay for data to appear on the slave side will
- 269 # be reduced but more bandwidth will be used for replication.
- 270 #
- 271 # By default we optimize for low latency, but in very high traffic conditions
- 272 # or when the master and slaves are many hops away, turning this to "yes" may
- 273 # be a good idea.
- 274 repl-disable-tcp-nodelay no
- 275
- 276 # Set the replication backlog size. The backlog is a buffer that accumulates
- 277 # slave data when slaves are disconnected for some time, so that when a slave
- 278 # wants to reconnect again, often a full resync is not needed, but a partial
- 279 # resync is enough, just passing the portion of data the slave missed while
- 280 # disconnected.
- 281 #
- 282 # The biggest the replication backlog, the longer the time the slave can be
- 283 # disconnected and later be able to perform a partial resynchronization.
- 284 #
- 285 # The backlog is only allocated once there is at least a slave connected.
- 286 #
- 287 # repl-backlog-size 1mb
- 288
- 289 # After a master has no longer connected slaves for some time, the backlog
- 290 # will be freed. The following option configures the amount of seconds that
- 291 # need to elapse, starting from the time the last slave disconnected, for
- 292 # the backlog buffer to be freed.
- 293 #
- 294 # A value of 0 means to never release the backlog.
- 295 #
- 296 # repl-backlog-ttl 3600
- 297
- 298 # The slave priority is an integer number published by Redis in the INFO output.
- 299 # It is used by Redis Sentinel in order to select a slave to promote into a
- 300 # master if the master is no longer working correctly.
- 301 #
- 302 # A slave with a low priority number is considered better for promotion, so
- 303 # for instance if there are three slaves with priority 10, 100, 25 Sentinel will
- 304 # pick the one with priority 10, that is the lowest.
- 305 #
- 306 # However a special priority of 0 marks the slave as not able to perform the
- 307 # role of master, so a slave with priority of 0 will never be selected by
- 308 # Redis Sentinel for promotion.
- 309 #
- 310 # By default the priority is 100.
- 311 slave-priority 100
- 312
- 313 # It is possible for a master to stop accepting writes if there are less than
- 314 # N slaves connected, having a lag less or equal than M seconds.
- 315 #
- 316 # The N slaves need to be in "online" state.
- 317 #
- 318 # The lag in seconds, that must be <= the specified value, is calculated from
- 319 # the last ping received from the slave, that is usually sent every second.
- 320 #
- 321 # This option does not GUARANTEES that N replicas will accept the write, but
- 322 # will limit the window of exposure for lost writes in case not enough slaves
- 323 # are available, to the specified number of seconds.
- 324 #
- 325 # For example to require at least 3 slaves with a lag <= 10 seconds use:
- 326 #
- 327 # min-slaves-to-write 3
- 328 # min-slaves-max-lag 10
- 329 #
- 330 # Setting one or the other to 0 disables the feature.
- 331 #
- 332 # By default min-slaves-to-write is set to 0 (feature disabled) and
- 333 # min-slaves-max-lag is set to 10.
- 334
- 335 ################################## SECURITY ###################################
- 336
- 337 # Require clients to issue AUTH <PASSWORD> before processing any other
- 338 # commands. This might be useful in environments in which you do not trust
- 339 # others with access to the host running redis-server.
- 340 #
- 341 # This should stay commented out for backward compatibility and because most
- 342 # people do not need auth (e.g. they run their own servers).
- 343 #
- 344 # Warning: since Redis is pretty fast an outside user can try up to
- 345 # 150k passwords per second against a good box. This means that you should
- 346 # use a very strong password otherwise it will be very easy to break.
- 347 #
- 348 # requirepass foobared
- 349
- 350 # Command renaming.
- 351 #
- 352 # It is possible to change the name of dangerous commands in a shared
- 353 # environment. For instance the CONFIG command may be renamed into something
- 354 # hard to guess so that it will still be available for internal-use tools
- 355 # but not available for general clients.
- 356 #
- 357 # Example:
- 358 #
- 359 # rename-command CONFIG b840fc02d524045429941cc15f59e41cb7be6c52
- 360 #
- 361 # It is also possible to completely kill a command by renaming it into
- 362 # an empty string:
- 363 #
- 364 # rename-command CONFIG ""
- 365 #
- 366 # Please note that changing the name of commands that are logged into the
- 367 # AOF file or transmitted to slaves may cause problems.
- 368
- 369 ################################### LIMITS ####################################
- 370
- 371 # Set the max number of connected clients at the same time. By default
- 372 # this limit is set to 10000 clients, however if the Redis server is not
- 373 # able to configure the process file limit to allow for the specified limit
- 374 # the max number of allowed clients is set to the current file limit
- 375 # minus 32 (as Redis reserves a few file descriptors for internal uses).
- 376 #
- 377 # Once the limit is reached Redis will close all the new connections sending
- 378 # an error 'max number of clients reached'.
- 379 #
- 380 # maxclients 10000
- 381
- 382 # Don't use more memory than the specified amount of bytes.
- 383 # When the memory limit is reached Redis will try to remove keys
- 384 # according to the eviction policy selected (see maxmemory-policy).
- 385 #
- 386 # If Redis can't remove keys according to the policy, or if the policy is
- 387 # set to 'noeviction', Redis will start to reply with errors to commands
- 388 # that would use more memory, like SET, LPUSH, and so on, and will continue
- 389 # to reply to read-only commands like GET.
- 390 #
- 391 # This option is usually useful when using Redis as an LRU cache, or to set
- 392 # a hard memory limit for an instance (using the 'noeviction' policy).
- 393 #
- 394 # WARNING: If you have slaves attached to an instance with maxmemory on,
- 395 # the size of the output buffers needed to feed the slaves are subtracted
- 396 # from the used memory count, so that network problems / resyncs will
- 397 # not trigger a loop where keys are evicted, and in turn the output
- 398 # buffer of slaves is full with DELs of keys evicted triggering the deletion
- 399 # of more keys, and so forth until the database is completely emptied.
- 400 #
- 401 # In short... if you have slaves attached it is suggested that you set a lower
- 402 # limit for maxmemory so that there is some free RAM on the system for slave
- 403 # output buffers (but this is not needed if the policy is 'noeviction').
- 404 #
- 405 # maxmemory <bytes>
- 406
- 407 # MAXMEMORY POLICY: how Redis will select what to remove when maxmemory
- 408 # is reached. You can select among five behaviors:
- 409 #
- 410 # volatile-lru -> remove the key with an expire set using an LRU algorithm
- 411 # allkeys-lru -> remove any key accordingly to the LRU algorithm
- 412 # volatile-random -> remove a random key with an expire set
- 413 # allkeys-random -> remove a random key, any key
- 414 # volatile-ttl -> remove the key with the nearest expire time (minor TTL)
- 415 # noeviction -> don't expire at all, just return an error on write operations
- 416 #
- 417 # Note: with any of the above policies, Redis will return an error on write
- 418 # operations, when there are not suitable keys for eviction.
- 419 #
- 420 # At the date of writing this commands are: set setnx setex append
- 421 # incr decr rpush lpush rpushx lpushx linsert lset rpoplpush sadd
- 422 # sinter sinterstore sunion sunionstore sdiff sdiffstore zadd zincrby
- 423 # zunionstore zinterstore hset hsetnx hmset hincrby incrby decrby
- 424 # getset mset msetnx exec sort
- 425 #
- 426 # The default is:
- 427 #
- 428 # maxmemory-policy volatile-lru
- 429
- 430 # LRU and minimal TTL algorithms are not precise algorithms but approximated
- 431 # algorithms (in order to save memory), so you can select as well the sample
- 432 # size to check. For instance for default Redis will check three keys and
- 433 # pick the one that was used less recently, you can change the sample size
- 434 # using the following configuration directive.
- 435 #
- 436 # maxmemory-samples 3
- 437
- 438 ############################## APPEND ONLY MODE ###############################
- 439
- 440 # By default Redis asynchronously dumps the dataset on disk. This mode is
- 441 # good enough in many applications, but an issue with the Redis process or
- 442 # a power outage may result into a few minutes of writes lost (depending on
- 443 # the configured save points).
- 444 #
- 445 # The Append Only File is an alternative persistence mode that provides
- 446 # much better durability. For instance using the default data fsync policy
- 447 # (see later in the config file) Redis can lose just one second of writes in a
- 448 # dramatic event like a server power outage, or a single write if something
- 449 # wrong with the Redis process itself happens, but the operating system is
- 450 # still running correctly.
- 451 #
- 452 # AOF and RDB persistence can be enabled at the same time without problems.
- 453 # If the AOF is enabled on startup Redis will load the AOF, that is the file
- 454 # with the better durability guarantees.
- 455 #
- 456 # Please check http://redis.io/topics/persistence for more information.
- 457
- 458 appendonly yes
- 459
- 460 # The name of the append only file (default: "appendonly.aof")
- 461
- 462 appendfilename "appendonly_6391.aof"
- 463
- 464 # The fsync() call tells the Operating System to actually write data on disk
- 465 # instead to wait for more data in the output buffer. Some OS will really flush
- 466 # data on disk, some other OS will just try to do it ASAP.
- 467 #
- 468 # Redis supports three different modes:
- 469 #
- 470 # no: don't fsync, just let the OS flush the data when it wants. Faster.
- 471 # always: fsync after every write to the append only log . Slow, Safest.
- 472 # everysec: fsync only one time every second. Compromise.
- 473 #
- 474 # The default is "everysec", as that's usually the right compromise between
- 475 # speed and data safety. It's up to you to understand if you can relax this to
- 476 # "no" that will let the operating system flush the output buffer when
- 477 # it wants, for better performances (but if you can live with the idea of
- 478 # some data loss consider the default persistence mode that's snapshotting),
- 479 # or on the contrary, use "always" that's very slow but a bit safer than
- 480 # everysec.
- 481 #
- 482 # More details please check the following article:
- 483 # http://antirez.com/post/redis-persistence-demystified.html
- 484 #
- 485 # If unsure, use "everysec".
- 486
- 487 # appendfsync always
- 488 appendfsync everysec
- 489 # appendfsync no
- 490
- 491 # When the AOF fsync policy is set to always or everysec, and a background
- 492 # saving process (a background save or AOF log background rewriting) is
- 493 # performing a lot of I/O against the disk, in some Linux configurations
- 494 # Redis may block too long on the fsync() call. Note that there is no fix for
- 495 # this currently, as even performing fsync in a different thread will block
- 496 # our synchronous write(2) call.
- 497 #
- 498 # In order to mitigate this problem it's possible to use the following option
- 499 # that will prevent fsync() from being called in the main process while a
- 500 # BGSAVE or BGREWRITEAOF is in progress.
- 501 #
- 502 # This means that while another child is saving, the durability of Redis is
- 503 # the same as "appendfsync none". In practical terms, this means that it is
- 504 # possible to lose up to 30 seconds of log in the worst scenario (with the
- 505 # default Linux settings).
- 506 #
- 507 # If you have latency problems turn this to "yes". Otherwise leave it as
- 508 # "no" that is the safest pick from the point of view of durability.
- 509
- 510 no-appendfsync-on-rewrite no
- 511
- 512 # Automatic rewrite of the append only file.
- 513 # Redis is able to automatically rewrite the log file implicitly calling
- 514 # BGREWRITEAOF when the AOF log size grows by the specified percentage.
- 515 #
- 516 # This is how it works: Redis remembers the size of the AOF file after the
- 517 # latest rewrite (if no rewrite has happened since the restart, the size of
- 518 # the AOF at startup is used).
- 519 #
- 520 # This base size is compared to the current size. If the current size is
- 521 # bigger than the specified percentage, the rewrite is triggered. Also
- 522 # you need to specify a minimal size for the AOF file to be rewritten, this
- 523 # is useful to avoid rewriting the AOF file even if the percentage increase
- 524 # is reached but it is still pretty small.
- 525 #
- 526 # Specify a percentage of zero in order to disable the automatic AOF
- 527 # rewrite feature.
- 528
- 529 auto-aof-rewrite-percentage 100
- 530 auto-aof-rewrite-min-size 64mb
- 531
- 532 # An AOF file may be found to be truncated at the end during the Redis
- 533 # startup process, when the AOF data gets loaded back into memory.
- 534 # This may happen when the system where Redis is running
- 535 # crashes, especially when an ext4 filesystem is mounted without the
- 536 # data=ordered option (however this can't happen when Redis itself
- 537 # crashes or aborts but the operating system still works correctly).
- 538 #
- 539 # Redis can either exit with an error when this happens, or load as much
- 540 # data as possible (the default now) and start if the AOF file is found
- 541 # to be truncated at the end. The following option controls this behavior.
- 542 #
- 543 # If aof-load-truncated is set to yes, a truncated AOF file is loaded and
- 544 # the Redis server starts emitting a log to inform the user of the event.
- 545 # Otherwise if the option is set to no, the server aborts with an error
- 546 # and refuses to start. When the option is set to no, the user requires
- 547 # to fix the AOF file using the "redis-check-aof" utility before to restart
- 548 # the server.
- 549 #
- 550 # Note that if the AOF file will be found to be corrupted in the middle
- 551 # the server will still exit with an error. This option only applies when
- 552 # Redis will try to read more data from the AOF file but not enough bytes
- 553 # will be found.
- 554 aof-load-truncated yes
- 555
- 556 ################################ LUA SCRIPTING ###############################
- 557
- 558 # Max execution time of a Lua script in milliseconds.
- 559 #
- 560 # If the maximum execution time is reached Redis will log that a script is
- 561 # still in execution after the maximum allowed time and will start to
- 562 # reply to queries with an error.
- 563 #
- 564 # When a long running script exceed the maximum execution time only the
- 565 # SCRIPT KILL and SHUTDOWN NOSAVE commands are available. The first can be
- 566 # used to stop a script that did not yet called write commands. The second
- 567 # is the only way to shut down the server in the case a write commands was
- 568 # already issue by the script but the user don't want to wait for the natural
- 569 # termination of the script.
- 570 #
- 571 # Set it to 0 or a negative value for unlimited execution without warnings.
- 572 lua-time-limit 5000
- 573
- 574 ################################## SLOW LOG ###################################
- 575
- 576 # The Redis Slow Log is a system to log queries that exceeded a specified
- 577 # execution time. The execution time does not include the I/O operations
- 578 # like talking with the client, sending the reply and so forth,
- 579 # but just the time needed to actually execute the command (this is the only
- 580 # stage of command execution where the thread is blocked and can not serve
- 581 # other requests in the meantime).
- 582 #
- 583 # You can configure the slow log with two parameters: one tells Redis
- 584 # what is the execution time, in microseconds, to exceed in order for the
- 585 # command to get logged, and the other parameter is the length of the
- 586 # slow log. When a new command is logged the oldest one is removed from the
- 587 # queue of logged commands.
- 588
- 589 # The following time is expressed in microseconds, so 1000000 is equivalent
- 590 # to one second. Note that a negative number disables the slow log, while
- 591 # a value of zero forces the logging of every command.
- 592 slowlog-log-slower-than 10000
- 593
- 594 # There is no limit to this length. Just be aware that it will consume memory.
- 595 # You can reclaim memory used by the slow log with SLOWLOG RESET.
- 596 slowlog-max-len 128
- 597
- 598 ################################ LATENCY MONITOR ##############################
- 599
- 600 # The Redis latency monitoring subsystem samples different operations
- 601 # at runtime in order to collect data related to possible sources of
- 602 # latency of a Redis instance.
- 603 #
- 604 # Via the LATENCY command this information is available to the user that can
- 605 # print graphs and obtain reports.
- 606 #
- 607 # The system only logs operations that were performed in a time equal or
- 608 # greater than the amount of milliseconds specified via the
- 609 # latency-monitor-threshold configuration directive. When its value is set
- 610 # to zero, the latency monitor is turned off.
- 611 #
- 612 # By default latency monitoring is disabled since it is mostly not needed
- 613 # if you don't have latency issues, and collecting data has a performance
- 614 # impact, that while very small, can be measured under big load. Latency
- 615 # monitoring can easily be enalbed at runtime using the command
- 616 # "CONFIG SET latency-monitor-threshold <milliseconds>" if needed.
- 617 latency-monitor-threshold 0
- 618
- 619 ############################# Event notification ##############################
- 620
- 621 # Redis can notify Pub/Sub clients about events happening in the key space.
- 622 # This feature is documented at http://redis.io/topics/notifications
- 623 #
- 624 # For instance if keyspace events notification is enabled, and a client
- 625 # performs a DEL operation on key "foo" stored in the Database 0, two
- 626 # messages will be published via Pub/Sub:
- 627 #
- 628 # PUBLISH __keyspace@0__:foo del
- 629 # PUBLISH __keyevent@0__:del foo
- 630 #
- 631 # It is possible to select the events that Redis will notify among a set
- 632 # of classes. Every class is identified by a single character:
- 633 #
- 634 # K Keyspace events, published with __keyspace@<db>__ prefix.
- 635 # E Keyevent events, published with __keyevent@<db>__ prefix.
- 636 # g Generic commands (non-type specific) like DEL, EXPIRE, RENAME, ...
- 637 # $ String commands
- 638 # l List commands
- 639 # s Set commands
- 640 # h Hash commands
- 641 # z Sorted set commands
- 642 # x Expired events (events generated every time a key expires)
- 643 # e Evicted events (events generated when a key is evicted for maxmemory)
- 644 # A Alias for g$lshzxe, so that the "AKE" string means all the events.
- 645 #
- 646 # The "notify-keyspace-events" takes as argument a string that is composed
- 647 # by zero or multiple characters. The empty string means that notifications
- 648 # are disabled at all.
- 649 #
- 650 # Example: to enable list and generic events, from the point of view of the
- 651 # event name, use:
- 652 #
- 653 # notify-keyspace-events Elg
- 654 #
- 655 # Example 2: to get the stream of the expired keys subscribing to channel
- 656 # name __keyevent@0__:expired use:
- 657 #
- 658 # notify-keyspace-events Ex
- 659 #
- 660 # By default all notifications are disabled because most users don't need
- 661 # this feature and the feature has some overhead. Note that if you don't
- 662 # specify at least one of K or E, no events will be delivered.
- 663 notify-keyspace-events ""
- 664
- 665 ############################### ADVANCED CONFIG ###############################
- 666
- 667 # Hashes are encoded using a memory efficient data structure when they have a
- 668 # small number of entries, and the biggest entry does not exceed a given
- 669 # threshold. These thresholds can be configured using the following directives.
- 670 hash-max-ziplist-entries 512
- 671 hash-max-ziplist-value 64
- 672
- 673 # Similarly to hashes, small lists are also encoded in a special way in order
- 674 # to save a lot of space. The special representation is only used when
- 675 # you are under the following limits:
- 676 list-max-ziplist-entries 512
- 677 list-max-ziplist-value 64
- 678
- 679 # Sets have a special encoding in just one case: when a set is composed
- 680 # of just strings that happens to be integers in radix 10 in the range
- 681 # of 64 bit signed integers.
- 682 # The following configuration setting sets the limit in the size of the
- 683 # set in order to use this special memory saving encoding.
- 684 set-max-intset-entries 512
- 685
- 686 # Similarly to hashes and lists, sorted sets are also specially encoded in
- 687 # order to save a lot of space. This encoding is only used when the length and
- 688 # elements of a sorted set are below the following limits:
- 689 zset-max-ziplist-entries 128
- 690 zset-max-ziplist-value 64
- 691
- 692 # HyperLogLog sparse representation bytes limit. The limit includes the
- 693 # 16 bytes header. When an HyperLogLog using the sparse representation crosses
- 694 # this limit, it is converted into the dense representation.
- 695 #
- 696 # A value greater than 16000 is totally useless, since at that point the
- 697 # dense representation is more memory efficient.
- 698 #
- 699 # The suggested value is ~ 3000 in order to have the benefits of
- 700 # the space efficient encoding without slowing down too much PFADD,
- 701 # which is O(N) with the sparse encoding. The value can be raised to
- 702 # ~ 10000 when CPU is not a concern, but space is, and the data set is
- 703 # composed of many HyperLogLogs with cardinality in the 0 - 15000 range.
- 704 hll-sparse-max-bytes 3000
- 705
- 706 # Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in
- 707 # order to help rehashing the main Redis hash table (the one mapping top-level
- 708 # keys to values). The hash table implementation Redis uses (see dict.c)
- 709 # performs a lazy rehashing: the more operation you run into a hash table
- 710 # that is rehashing, the more rehashing "steps" are performed, so if the
- 711 # server is idle the rehashing is never complete and some more memory is used
- 712 # by the hash table.
- 713 #
- 714 # The default is to use this millisecond 10 times every second in order to
- 715 # active rehashing the main dictionaries, freeing memory when possible.
- 716 #
- 717 # If unsure:
- 718 # use "activerehashing no" if you have hard latency requirements and it is
- 719 # not a good thing in your environment that Redis can reply form time to time
- 720 # to queries with 2 milliseconds delay.
- 721 #
- 722 # use "activerehashing yes" if you don't have such hard requirements but
- 723 # want to free memory asap when possible.
- 724 activerehashing yes
- 725
- 726 # The client output buffer limits can be used to force disconnection of clients
- 727 # that are not reading data from the server fast enough for some reason (a
- 728 # common reason is that a Pub/Sub client can't consume messages as fast as the
- 729 # publisher can produce them).
- 730 #
- 731 # The limit can be set differently for the three different classes of clients:
- 732 #
- 733 # normal -> normal clients including MONITOR clients
- 734 # slave -> slave clients
- 735 # pubsub -> clients subscribed to at least one pubsub channel or pattern
- 736 #
- 737 # The syntax of every client-output-buffer-limit directive is the following:
- 738 #
- 739 # client-output-buffer-limit <class> <hard limit> <soft limit> <soft seconds>
- 740 #
- 741 # A client is immediately disconnected once the hard limit is reached, or if
- 742 # the soft limit is reached and remains reached for the specified number of
- 743 # seconds (continuously).
- 744 # So for instance if the hard limit is 32 megabytes and the soft limit is
- 745 # 16 megabytes / 10 seconds, the client will get disconnected immediately
- 746 # if the size of the output buffers reach 32 megabytes, but will also get
- 747 # disconnected if the client reaches 16 megabytes and continuously overcomes
- 748 # the limit for 10 seconds.
- 749 #
- 750 # By default normal clients are not limited because they don't receive data
- 751 # without asking (in a push way), but just after a request, so only
- 752 # asynchronous clients may create a scenario where data is requested faster
- 753 # than it can read.
- 754 #
- 755 # Instead there is a default limit for pubsub and slave clients, since
- 756 # subscribers and slaves receive data in a push fashion.
- 757 #
- 758 # Both the hard or the soft limit can be disabled by setting them to zero.
- 759 client-output-buffer-limit normal 0 0 0
- 760 client-output-buffer-limit slave 256mb 64mb 60
- 761 client-output-buffer-limit pubsub 32mb 8mb 60
- 762
- 763 # Redis calls an internal function to perform many background tasks, like
- 764 # closing connections of clients in timeout, purging expired keys that are
- 765 # never requested, and so forth.
- 766 #
- 767 # Not all tasks are performed with the same frequency, but Redis checks for
- 768 # tasks to perform accordingly to the specified "hz" value.
- 769 #
- 770 # By default "hz" is set to 10. Raising the value will use more CPU when
- 771 # Redis is idle, but at the same time will make Redis more responsive when
- 772 # there are many keys expiring at the same time, and timeouts may be
- 773 # handled with more precision.
- 774 #
- 775 # The range is between 1 and 500, however a value over 100 is usually not
- 776 # a good idea. Most users should use the default of 10 and raise this up to
- 777 # 100 only in environments where very low latency is required.
- 778 hz 10
- 779
- 780 # When a child rewrites the AOF file, if the following option is enabled
- 781 # the file will be fsync-ed every 32 MB of data generated. This is useful
- 782 # in order to commit the file to the disk more incrementally and avoid
- 783 # big latency spikes.
- 784 aof-rewrite-incremental-fsync yes
redis_6391.conf源码
(7) 重启指定端口的服务,例如此处在Redis按照src目录下,运行./redis-server redis_6391.conf即可启动服务,待服务完成启动成功后,即可把指定的AOF文件数据加载进去(PS:此步骤需要先确认指定目录下的AOF文件已被替换成目标AOF文件,期间可以多次重启实现具体AOF文件加载)
以下给出我本机使用Redis加载启动大小为1.7G的aof文件,由于文件比较大,所以加载的时间有点长,此处是加载了60秒。
liuzhen@liuzhen-ubuntu:~/redis-2.8.17/src$ ./redis-server redis_6391.conf [68180] 19 Jul 15:02:07.997 * Increased maximum number of open files to 10032 (it was originally set to 1024). _._ _.-``__ ''-._ _.-`` `. `_. ''-._ Redis 2.8.17 (00000000/0) 64 bit .-`` .-```. ```\/ _.,_ ''-._ ( ' , .-` | `, ) Running in stand alone mode |`-._`-...-` __...-.``-._|'` _.-'| Port: 6391 | `-._ `._ / _.-' | PID: 68180 `-._ `-._ `-./ _.-' _.-' |`-._`-._ `-.__.-' _.-'_.-'| | `-._`-._ _.-'_.-' | http://redis.io `-._ `-._`-.__.-'_.-' _.-' |`-._`-._ `-.__.-' _.-'_.-'| | `-._`-._ _.-'_.-' | `-._ `-._`-.__.-'_.-' _.-' `-._ `-.__.-' _.-' `-._ _.-' `-.__.-' [68180] 19 Jul 15:02:08.011 # Server started, Redis version 2.8.17 [68180] 19 Jul 15:05:12.843 * DB loaded from append only file: 184.831 seconds [68180] 19 Jul 15:05:12.843 * The server is now ready to accept connections on port 6391 [68180] 19 Jul 15:05:13.008 * 10000 changes in 60 seconds. Saving... [68180] 19 Jul 15:05:13.084 * Background saving started by pid 68228 [68228] 19 Jul 15:05:47.548 * DB saved on disk [68228] 19 Jul 15:05:47.613 * RDB: 23 MB of memory used by copy-on-write [68180] 19 Jul 15:05:47.717 * Background saving terminated with success [68180] 19 Jul 15:07:54.064 * DB saved on disk [68180] 19 Jul 15:08:58.096 * Asynchronous AOF fsync is taking too long (disk is busy?). Writing the AOF buffer without waiting for fsync to complete, this may slow down Redis. [68180] 19 Jul 16:49:14.515 * Background saving started by pid 90980 [90980] 19 Jul 16:56:56.883 * DB saved on disk [90980] 19 Jul 16:56:56.966 * RDB: 4 MB of memory used by copy-on-write [68180] 19 Jul 16:56:57.418 * Background saving terminated with success
(8)打开Redis客户端,运行./redis-cli -p 6391,客户端启动成功后,运行命令save,等待命令运行成功后,即可得到本步骤最终目标的RDB持久化文件(PS:此处如果是在线上环境尝试,建议采用bgsave命令)
此处给出,使用AOF文件还原数据后,查看具体数据信息的结果:
liuzhen@liuzhen-ubuntu:~/redis-2.8.17/src$ ./redis-cli -p 6391 127.0.0.1:6391> info # Server redis_version:2.8.17 redis_git_sha1:00000000 redis_git_dirty:0 redis_build_id:4ba260b6ab802599 redis_mode:standalone os:Linux 4.13.0-39-generic x86_64 arch_bits:64 multiplexing_api:epoll gcc_version:5.4.0 process_id:68180 run_id:97cddc494e3924885bacb03776dfe09e8fa055f9 tcp_port:6391 uptime_in_seconds:9400 uptime_in_days:0 hz:10 lru_clock:5266472 config_file:/home/liuzhen/redis-2.8.17/src/redis_6391.conf # Clients connected_clients:1 client_longest_output_list:0 client_biggest_input_buf:0 blocked_clients:0 # Memory used_memory:2239514040 used_memory_human:2.09G used_memory_rss:330895360 used_memory_peak:2272377648 used_memory_peak_human:2.12G used_memory_lua:38912 mem_fragmentation_ratio:0.15 mem_allocator:jemalloc-3.6.0 # Persistence loading:0 rdb_changes_since_last_save:0 rdb_bgsave_in_progress:0 rdb_last_save_time:1531990617 rdb_last_bgsave_status:ok rdb_last_bgsave_time_sec:463 rdb_current_bgsave_time_sec:-1 aof_enabled:1 aof_rewrite_in_progress:0 aof_rewrite_scheduled:0 aof_last_rewrite_time_sec:-1 aof_current_rewrite_time_sec:-1 aof_last_bgrewrite_status:ok aof_last_write_status:ok aof_current_size:1700508277 aof_base_size:1699947297 aof_pending_rewrite:0 aof_buffer_length:0 aof_rewrite_buffer_length:0 aof_pending_bio_fsync:0 aof_delayed_fsync:1 # Stats total_connections_received:2 total_commands_processed:281 instantaneous_ops_per_sec:0 rejected_connections:0 sync_full:0 sync_partial_ok:0 sync_partial_err:0 expired_keys:9290 evicted_keys:0 keyspace_hits:1065050 keyspace_misses:0 pubsub_channels:0 pubsub_patterns:0 latest_fork_usec:101807 # Replication role:master connected_slaves:0 master_repl_offset:0 repl_backlog_active:0 repl_backlog_size:1048576 repl_backlog_first_byte_offset:0 repl_backlog_histlen:0 # CPU used_cpu_sys:46.01 used_cpu_user:189.71 used_cpu_sys_children:134.11 used_cpu_user_children:79.12 # Keyspace db1:keys=1146336,expires=51965,avg_ttl=276142509 127.0.0.1:6391>
备注:在Redis指定端口服务加载给定的AOF文件时,如果AOF文件过大,系统可能会报如下错误:
Can’t save in background: fork: Cannot allocate memory
解决办法:
修改系统/etc/sysctl.conf文件,并添加以下内容:
vm.overcommit_memory=1
在 FreeBSD上:
sudo /etc/rc.d/sysctl reload
在 Linux上:
sudo sysctl -p /etc/sysctl.conf
参考资料:
- 1.redis中的数据快照、AOF、数据恢复、主从复制介绍及使用(https://blog.csdn.net/zhu_xun/article/details/16806697)
- 2.redis如何利用appendonly.aof恢复数据(https://blog.csdn.net/sanbingyutuoniao123/article/details/50484674)
- 3.Redis持久存储(AOF/Snapshot)(http://shift-alt-ctrl.iteye.com/blog/1878716)
- 4.redis dump.rdb appendonly.aof 文件路径修改(https://blog.csdn.net/yangxujia/article/details/51010222)
- 5.超大数据快速导入MySQL(https://blog.csdn.net/dingding_12345/article/details/78646484)
- 6.redis Can’t save in background: fork: Cannot allocate memory 解决及原理(https://blog.csdn.net/zqz_zqz/article/details/53384854)
- 7.redis bgsave failed because fork Cannot allocate memory(https://stackoverflow.com/questions/11752544/redis-bgsave-failed-because-fork-cannot-allocate-memory)