• INCLUDES
  • MODULES
  • NETWORK
  • GENERAL
  • SNAPSHOTTING
  • REPLICATION
  • SECURITY
  • CLIENTS
  • MEMORY MANAGEMENT
  • LAZY FREEING
  • APPEND ONLY MODE
  • LUA SCRIPTING
  • REDIS CLUSTER
  • CLUSTER DOCKER/NAT support
  • SLOW LOG
  • LATENCY MONITOR
  • EVENT NOTIFICATION
  • ADVANCED CONFIG
  • ACTIVE DEFRAGMENTATION

    本文来自redis 官方配置文件

    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
    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.

    INCLUDES

    1. ################################## INCLUDES ###################################
    3. # Include one or more other config files here.  This is useful if you
    4. # have a standard template that goes to all Redis servers but also need
    5. # to customize a few per-server settings.  Include files can include
    6. # other files, so use this wisely.
    7. #
    8. # Notice option "include" won't be rewritten by command "CONFIG REWRITE"
    9. # from admin or Redis Sentinel. Since Redis always uses the last processed
    10. # line as value of a configuration directive, you'd better put includes
    11. # at the beginning of this file to avoid overwriting config change at runtime.
    12. #
    13. # If instead you are interested in using includes to override configuration
    14. # options, it is better to use include as the last line.
    15. #
    16. # include /path/to/local.conf
    17. # include /path/to/other.conf

    MODULES

    1. ################################## MODULES #####################################
    3. # Load modules at startup. If the server is not able to load modules
    4. # it will abort. It is possible to use multiple loadmodule directives.
    5. #
    6. # loadmodule /path/to/my_module.so
    7. # loadmodule /path/to/other_module.so

    NETWORK

    1. ################################## NETWORK #####################################
    3. # By default, if no "bind" configuration directive is specified, Redis listens
    4. # for connections from all the network interfaces available on the server.
    5. # It is possible to listen to just one or multiple selected interfaces using
    6. # the "bind" configuration directive, followed by one or more IP addresses.
    7. #
    8. # Examples:
    9. #
    10. # bind 192.168.1.100 10.0.0.1
    11. # bind 127.0.0.1 ::1
    12. #
    13. # ~~~ WARNING ~~~ If the computer running Redis is directly exposed to the
    14. # internet, binding to all the interfaces is dangerous and will expose the
    15. # instance to everybody on the internet. So by default we uncomment the
    16. # following bind directive, that will force Redis to listen only into
    17. # the IPv4 lookback interface address (this means Redis will be able to
    18. # accept connections only from clients running into the same computer it
    19. # is running).
    20. #
    21. # IF YOU ARE SURE YOU WANT YOUR INSTANCE TO LISTEN TO ALL THE INTERFACES
    22. # JUST COMMENT THE FOLLOWING LINE.
    23. # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    24. bind 127.0.0.1
    26. # Protected mode is a layer of security protection, in order to avoid that
    27. # Redis instances left open on the internet are accessed and exploited.
    28. #
    29. # When protected mode is on and if:
    30. #
    31. # 1) The server is not binding explicitly to a set of addresses using the
    32. #    "bind" directive.
    33. # 2) No password is configured.
    34. #
    35. # The server only accepts connections from clients connecting from the
    36. # IPv4 and IPv6 loopback addresses 127.0.0.1 and ::1, and from Unix domain
    37. # sockets.
    38. #
    39. # By default protected mode is enabled. You should disable it only if
    40. # you are sure you want clients from other hosts to connect to Redis
    41. # even if no authentication is configured, nor a specific set of interfaces
    42. # are explicitly listed using the "bind" directive.
    43. protected-mode yes
    45. # Accept connections on the specified port, default is 6379 (IANA #815344).
    46. # If port 0 is specified Redis will not listen on a TCP socket.
    47. port 6379
    49. # TCP listen() backlog.
    50. #
    51. # In high requests-per-second environments you need an high backlog in order
    52. # to avoid slow clients connections issues. Note that the Linux kernel
    53. # will silently truncate it to the value of /proc/sys/net/core/somaxconn so
    54. # make sure to raise both the value of somaxconn and tcp_max_syn_backlog
    55. # in order to get the desired effect.
    56. tcp-backlog 511
    58. # Unix socket.
    59. #
    60. # Specify the path for the Unix socket that will be used to listen for
    61. # incoming connections. There is no default, so Redis will not listen
    62. # on a unix socket when not specified.
    63. #
    64. # unixsocket /tmp/redis.sock
    65. # unixsocketperm 700
    67. # Close the connection after a client is idle for N seconds (0 to disable)
    68. timeout 0
    70. # TCP keepalive.
    71. #
    72. # If non-zero, use SO_KEEPALIVE to send TCP ACKs to clients in absence
    73. # of communication. This is useful for two reasons:
    74. #
    75. # 1) Detect dead peers.
    76. # 2) Take the connection alive from the point of view of network
    77. #    equipment in the middle.
    78. #
    79. # On Linux, the specified value (in seconds) is the period used to send ACKs.
    80. # Note that to close the connection the double of the time is needed.
    81. # On other kernels the period depends on the kernel configuration.
    82. #
    83. # A reasonable value for this option is 300 seconds, which is the new
    84. # Redis default starting with Redis 3.2.1.
    85. tcp-keepalive 300

    GENERAL

    1. ################################# GENERAL #####################################
    3. # By default Redis does not run as a daemon. Use 'yes' if you need it.
    4. # Note that Redis will write a pid file in /var/run/redis.pid when daemonized.
    5. daemonize yes
    7. # If you run Redis from upstart or systemd, Redis can interact with your
    8. # supervision tree. Options:
    9. #   supervised no      - no supervision interaction
    10. #   supervised upstart - signal upstart by putting Redis into SIGSTOP mode
    11. #   supervised systemd - signal systemd by writing READY=1 to $NOTIFY_SOCKET
    12. #   supervised auto    - detect upstart or systemd method based on
    13. #                        UPSTART_JOB or NOTIFY_SOCKET environment variables
    14. # Note: these supervision methods only signal "process is ready."
    15. #       They do not enable continuous liveness pings back to your supervisor.
    16. supervised no
    18. # If a pid file is specified, Redis writes it where specified at startup
    19. # and removes it at exit.
    20. #
    21. # When the server runs non daemonized, no pid file is created if none is
    22. # specified in the configuration. When the server is daemonized, the pid file
    23. # is used even if not specified, defaulting to "/var/run/redis.pid".
    24. #
    25. # Creating a pid file is best effort: if Redis is not able to create it
    26. # nothing bad happens, the server will start and run normally.
    27. pidfile /var/run/redis_6379.pid
    29. # Specify the server verbosity level.
    30. # This can be one of:
    31. # debug (a lot of information, useful for development/testing)
    32. # verbose (many rarely useful info, but not a mess like the debug level)
    33. # notice (moderately verbose, what you want in production probably)
    34. # warning (only very important / critical messages are logged)
    35. loglevel notice
    37. # Specify the log file name. Also the empty string can be used to force
    38. # Redis to log on the standard output. Note that if you use standard
    39. # output for logging but daemonize, logs will be sent to /dev/null
    40. logfile ""
    42. # To enable logging to the system logger, just set 'syslog-enabled' to yes,
    43. # and optionally update the other syslog parameters to suit your needs.
    44. # syslog-enabled no
    46. # Specify the syslog identity.
    47. # syslog-ident redis
    49. # Specify the syslog facility. Must be USER or between LOCAL0-LOCAL7.
    50. # syslog-facility local0
    52. # Set the number of databases. The default database is DB 0, you can select
    53. # a different one on a per-connection basis using SELECT <dbid> where
    54. # dbid is a number between 0 and 'databases'-1
    55. databases 16
    57. # By default Redis shows an ASCII art logo only when started to log to the
    58. # standard output and if the standard output is a TTY. Basically this means
    59. # that normally a logo is displayed only in interactive sessions.
    60. #
    61. # However it is possible to force the pre-4.0 behavior and always show a
    62. # ASCII art logo in startup logs by setting the following option to yes.
    63. always-show-logo yes

    SNAPSHOTTING

    1. ################################ SNAPSHOTTING  ################################
    2. #
    3. # Save the DB on disk:
    4. #
    5. #   save <seconds> <changes>
    6. #
    7. #   Will save the DB if both the given number of seconds and the given
    8. #   number of write operations against the DB occurred.
    9. #
    10. #   In the example below the behaviour will be to save:
    11. #   after 900 sec (15 min) if at least 1 key changed
    12. #   after 300 sec (5 min) if at least 10 keys changed
    13. #   after 60 sec if at least 10000 keys changed
    14. #
    15. #   Note: you can disable saving completely by commenting out all "save" lines.
    16. #
    17. #   It is also possible to remove all the previously configured save
    18. #   points by adding a save directive with a single empty string argument
    19. #   like in the following example:
    20. #
    21. #   save ""
    23. save 900 1
    24. save 300 10
    25. save 60 10000
    27. # By default Redis will stop accepting writes if RDB snapshots are enabled
    28. # (at least one save point) and the latest background save failed.
    29. # This will make the user aware (in a hard way) that data is not persisting
    30. # on disk properly, otherwise chances are that no one will notice and some
    31. # disaster will happen.
    32. #
    33. # If the background saving process will start working again Redis will
    34. # automatically allow writes again.
    35. #
    36. # However if you have setup your proper monitoring of the Redis server
    37. # and persistence, you may want to disable this feature so that Redis will
    38. # continue to work as usual even if there are problems with disk,
    39. # permissions, and so forth.
    40. stop-writes-on-bgsave-error yes
    42. # Compress string objects using LZF when dump .rdb databases?
    43. # For default that's set to 'yes' as it's almost always a win.
    44. # If you want to save some CPU in the saving child set it to 'no' but
    45. # the dataset will likely be bigger if you have compressible values or keys.
    46. rdbcompression yes
    48. # Since version 5 of RDB a CRC64 checksum is placed at the end of the file.
    49. # This makes the format more resistant to corruption but there is a performance
    50. # hit to pay (around 10%) when saving and loading RDB files, so you can disable it
    51. # for maximum performances.
    52. #
    53. # RDB files created with checksum disabled have a checksum of zero that will
    54. # tell the loading code to skip the check.
    55. rdbchecksum yes
    57. # The filename where to dump the DB
    58. dbfilename dump.rdb
    60. # The working directory.
    61. #
    62. # The DB will be written inside this directory, with the filename specified
    63. # above using the 'dbfilename' configuration directive.
    64. #
    65. # The Append Only File will also be created inside this directory.
    66. #
    67. # Note that you must specify a directory here, not a file name.
    68. dir ./

    REPLICATION

    1. ################################# REPLICATION #################################
    3. # Master-Slave replication. Use slaveof to make a Redis instance a copy of
    4. # another Redis server. A few things to understand ASAP about Redis replication.
    5. #
    6. # 1) Redis replication is asynchronous, but you can configure a master to
    7. #    stop accepting writes if it appears to be not connected with at least
    8. #    a given number of slaves.
    9. # 2) Redis slaves are able to perform a partial resynchronization with the
    10. #    master if the replication link is lost for a relatively small amount of
    11. #    time. You may want to configure the replication backlog size (see the next
    12. #    sections of this file) with a sensible value depending on your needs.
    13. # 3) Replication is automatic and does not need user intervention. After a
    14. #    network partition slaves automatically try to reconnect to masters
    15. #    and resynchronize with them.
    16. #
    17. # slaveof <masterip> <masterport>
    19. # If the master is password protected (using the "requirepass" configuration
    20. # directive below) it is possible to tell the slave to authenticate before
    21. # starting the replication synchronization process, otherwise the master will
    22. # refuse the slave request.
    23. #
    24. # masterauth <master-password>
    26. # When a slave loses its connection with the master, or when the replication
    27. # is still in progress, the slave can act in two different ways:
    28. #
    29. # 1) if slave-serve-stale-data is set to 'yes' (the default) the slave will
    30. #    still reply to client requests, possibly with out of date data, or the
    31. #    data set may just be empty if this is the first synchronization.
    32. #
    33. # 2) if slave-serve-stale-data is set to 'no' the slave will reply with
    34. #    an error "SYNC with master in progress" to all the kind of commands
    35. #    but to INFO and SLAVEOF.
    36. #
    37. slave-serve-stale-data yes
    39. # You can configure a slave instance to accept writes or not. Writing against
    40. # a slave instance may be useful to store some ephemeral data (because data
    41. # written on a slave will be easily deleted after resync with the master) but
    42. # may also cause problems if clients are writing to it because of a
    43. # misconfiguration.
    44. #
    45. # Since Redis 2.6 by default slaves are read-only.
    46. #
    47. # Note: read only slaves are not designed to be exposed to untrusted clients
    48. # on the internet. It's just a protection layer against misuse of the instance.
    49. # Still a read only slave exports by default all the administrative commands
    50. # such as CONFIG, DEBUG, and so forth. To a limited extent you can improve
    51. # security of read only slaves using 'rename-command' to shadow all the
    52. # administrative / dangerous commands.
    53. slave-read-only yes
    55. # Replication SYNC strategy: disk or socket.
    56. #
    57. # -------------------------------------------------------
    58. # WARNING: DISKLESS REPLICATION IS EXPERIMENTAL CURRENTLY
    59. # -------------------------------------------------------
    60. #
    61. # New slaves and reconnecting slaves that are not able to continue the replication
    62. # process just receiving differences, need to do what is called a "full
    63. # synchronization". An RDB file is transmitted from the master to the slaves.
    64. # The transmission can happen in two different ways:
    65. #
    66. # 1) Disk-backed: The Redis master creates a new process that writes the RDB
    67. #                 file on disk. Later the file is transferred by the parent
    68. #                 process to the slaves incrementally.
    69. # 2) Diskless: The Redis master creates a new process that directly writes the
    70. #              RDB file to slave sockets, without touching the disk at all.
    71. #
    72. # With disk-backed replication, while the RDB file is generated, more slaves
    73. # can be queued and served with the RDB file as soon as the current child producing
    74. # the RDB file finishes its work. With diskless replication instead once
    75. # the transfer starts, new slaves arriving will be queued and a new transfer
    76. # will start when the current one terminates.
    77. #
    78. # When diskless replication is used, the master waits a configurable amount of
    79. # time (in seconds) before starting the transfer in the hope that multiple slaves
    80. # will arrive and the transfer can be parallelized.
    81. #
    82. # With slow disks and fast (large bandwidth) networks, diskless replication
    83. # works better.
    84. repl-diskless-sync no
    86. # When diskless replication is enabled, it is possible to configure the delay
    87. # the server waits in order to spawn the child that transfers the RDB via socket
    88. # to the slaves.
    89. #
    90. # This is important since once the transfer starts, it is not possible to serve
    91. # new slaves arriving, that will be queued for the next RDB transfer, so the server
    92. # waits a delay in order to let more slaves arrive.
    93. #
    94. # The delay is specified in seconds, and by default is 5 seconds. To disable
    95. # it entirely just set it to 0 seconds and the transfer will start ASAP.
    96. repl-diskless-sync-delay 5
    98. # Slaves send PINGs to server in a predefined interval. It's possible to change
    99. # this interval with the repl_ping_slave_period option. The default value is 10
    100. # seconds.
    101. #
    102. # repl-ping-slave-period 10
    104. # The following option sets the replication timeout for:
    105. #
    106. # 1) Bulk transfer I/O during SYNC, from the point of view of slave.
    107. # 2) Master timeout from the point of view of slaves (data, pings).
    108. # 3) Slave timeout from the point of view of masters (REPLCONF ACK pings).
    109. #
    110. # It is important to make sure that this value is greater than the value
    111. # specified for repl-ping-slave-period otherwise a timeout will be detected
    112. # every time there is low traffic between the master and the slave.
    113. #
    114. # repl-timeout 60
    116. # Disable TCP_NODELAY on the slave socket after SYNC?
    117. #
    118. # If you select "yes" Redis will use a smaller number of TCP packets and
    119. # less bandwidth to send data to slaves. But this can add a delay for
    120. # the data to appear on the slave side, up to 40 milliseconds with
    121. # Linux kernels using a default configuration.
    122. #
    123. # If you select "no" the delay for data to appear on the slave side will
    124. # be reduced but more bandwidth will be used for replication.
    125. #
    126. # By default we optimize for low latency, but in very high traffic conditions
    127. # or when the master and slaves are many hops away, turning this to "yes" may
    128. # be a good idea.
    129. repl-disable-tcp-nodelay no
    131. # Set the replication backlog size. The backlog is a buffer that accumulates
    132. # slave data when slaves are disconnected for some time, so that when a slave
    133. # wants to reconnect again, often a full resync is not needed, but a partial
    134. # resync is enough, just passing the portion of data the slave missed while
    135. # disconnected.
    136. #
    137. # The bigger the replication backlog, the longer the time the slave can be
    138. # disconnected and later be able to perform a partial resynchronization.
    139. #
    140. # The backlog is only allocated once there is at least a slave connected.
    141. #
    142. # repl-backlog-size 1mb
    144. # After a master has no longer connected slaves for some time, the backlog
    145. # will be freed. The following option configures the amount of seconds that
    146. # need to elapse, starting from the time the last slave disconnected, for
    147. # the backlog buffer to be freed.
    148. #
    149. # Note that slaves never free the backlog for timeout, since they may be
    150. # promoted to masters later, and should be able to correctly "partially
    151. # resynchronize" with the slaves: hence they should always accumulate backlog.
    152. #
    153. # A value of 0 means to never release the backlog.
    154. #
    155. # repl-backlog-ttl 3600
    157. # The slave priority is an integer number published by Redis in the INFO output.
    158. # It is used by Redis Sentinel in order to select a slave to promote into a
    159. # master if the master is no longer working correctly.
    160. #
    161. # A slave with a low priority number is considered better for promotion, so
    162. # for instance if there are three slaves with priority 10, 100, 25 Sentinel will
    163. # pick the one with priority 10, that is the lowest.
    164. #
    165. # However a special priority of 0 marks the slave as not able to perform the
    166. # role of master, so a slave with priority of 0 will never be selected by
    167. # Redis Sentinel for promotion.
    168. #
    169. # By default the priority is 100.
    170. slave-priority 100
    172. # It is possible for a master to stop accepting writes if there are less than
    173. # N slaves connected, having a lag less or equal than M seconds.
    174. #
    175. # The N slaves need to be in "online" state.
    176. #
    177. # The lag in seconds, that must be <= the specified value, is calculated from
    178. # the last ping received from the slave, that is usually sent every second.
    179. #
    180. # This option does not GUARANTEE that N replicas will accept the write, but
    181. # will limit the window of exposure for lost writes in case not enough slaves
    182. # are available, to the specified number of seconds.
    183. #
    184. # For example to require at least 3 slaves with a lag <= 10 seconds use:
    185. #
    186. # min-slaves-to-write 3
    187. # min-slaves-max-lag 10
    188. #
    189. # Setting one or the other to 0 disables the feature.
    190. #
    191. # By default min-slaves-to-write is set to 0 (feature disabled) and
    192. # min-slaves-max-lag is set to 10.
    194. # A Redis master is able to list the address and port of the attached
    195. # slaves in different ways. For example the "INFO replication" section
    196. # offers this information, which is used, among other tools, by
    197. # Redis Sentinel in order to discover slave instances.
    198. # Another place where this info is available is in the output of the
    199. # "ROLE" command of a master.
    200. #
    201. # The listed IP and address normally reported by a slave is obtained
    202. # in the following way:
    203. #
    204. #   IP: The address is auto detected by checking the peer address
    205. #   of the socket used by the slave to connect with the master.
    206. #
    207. #   Port: The port is communicated by the slave during the replication
    208. #   handshake, and is normally the port that the slave is using to
    209. #   list for connections.
    210. #
    211. # However when port forwarding or Network Address Translation (NAT) is
    212. # used, the slave may be actually reachable via different IP and port
    213. # pairs. The following two options can be used by a slave in order to
    214. # report to its master a specific set of IP and port, so that both INFO
    215. # and ROLE will report those values.
    216. #
    217. # There is no need to use both the options if you need to override just
    218. # the port or the IP address.
    219. #
    220. # slave-announce-ip 5.5.5.5
    221. # slave-announce-port 1234

    SECURITY

    1. ################################## SECURITY ###################################
    3. # Require clients to issue AUTH <PASSWORD> before processing any other
    4. # commands.  This might be useful in environments in which you do not trust
    5. # others with access to the host running redis-server.
    6. #
    7. # This should stay commented out for backward compatibility and because most
    8. # people do not need auth (e.g. they run their own servers).
    9. #
    10. # Warning: since Redis is pretty fast an outside user can try up to
    11. # 150k passwords per second against a good box. This means that you should
    12. # use a very strong password otherwise it will be very easy to break.
    13. #
    14. # requirepass foobared
    16. # Command renaming.
    17. #
    18. # It is possible to change the name of dangerous commands in a shared
    19. # environment. For instance the CONFIG command may be renamed into something
    20. # hard to guess so that it will still be available for internal-use tools
    21. # but not available for general clients.
    22. #
    23. # Example:
    24. #
    25. # rename-command CONFIG b840fc02d524045429941cc15f59e41cb7be6c52
    26. #
    27. # It is also possible to completely kill a command by renaming it into
    28. # an empty string:
    29. #
    30. # rename-command CONFIG ""
    31. #
    32. # Please note that changing the name of commands that are logged into the
    33. # AOF file or transmitted to slaves may cause problems.

    CLIENTS

    1. ################################### CLIENTS ####################################
    3. # Set the max number of connected clients at the same time. By default
    4. # this limit is set to 10000 clients, however if the Redis server is not
    5. # able to configure the process file limit to allow for the specified limit
    6. # the max number of allowed clients is set to the current file limit
    7. # minus 32 (as Redis reserves a few file descriptors for internal uses).
    8. #
    9. # Once the limit is reached Redis will close all the new connections sending
    10. # an error 'max number of clients reached'.
    11. #
    12. # maxclients 10000

    MEMORY MANAGEMENT

    1. ############################## MEMORY MANAGEMENT ################################
    3. # Set a memory usage limit to the specified amount of bytes.
    4. # When the memory limit is reached Redis will try to remove keys
    5. # according to the eviction policy selected (see maxmemory-policy).
    6. #
    7. # If Redis can't remove keys according to the policy, or if the policy is
    8. # set to 'noeviction', Redis will start to reply with errors to commands
    9. # that would use more memory, like SET, LPUSH, and so on, and will continue
    10. # to reply to read-only commands like GET.
    11. #
    12. # This option is usually useful when using Redis as an LRU or LFU cache, or to
    13. # set a hard memory limit for an instance (using the 'noeviction' policy).
    14. #
    15. # WARNING: If you have slaves attached to an instance with maxmemory on,
    16. # the size of the output buffers needed to feed the slaves are subtracted
    17. # from the used memory count, so that network problems / resyncs will
    18. # not trigger a loop where keys are evicted, and in turn the output
    19. # buffer of slaves is full with DELs of keys evicted triggering the deletion
    20. # of more keys, and so forth until the database is completely emptied.
    21. #
    22. # In short... if you have slaves attached it is suggested that you set a lower
    23. # limit for maxmemory so that there is some free RAM on the system for slave
    24. # output buffers (but this is not needed if the policy is 'noeviction').
    25. #
    26. # maxmemory <bytes>
    28. # MAXMEMORY POLICY: how Redis will select what to remove when maxmemory
    29. # is reached. You can select among five behaviors:
    30. #
    31. # volatile-lru -> Evict using approximated LRU among the keys with an expire set.
    32. # allkeys-lru -> Evict any key using approximated LRU.
    33. # volatile-lfu -> Evict using approximated LFU among the keys with an expire set.
    34. # allkeys-lfu -> Evict any key using approximated LFU.
    35. # volatile-random -> Remove a random key among the ones with an expire set.
    36. # allkeys-random -> Remove a random key, any key.
    37. # volatile-ttl -> Remove the key with the nearest expire time (minor TTL)
    38. # noeviction -> Don't evict anything, just return an error on write operations.
    39. #
    40. # LRU means Least Recently Used
    41. # LFU means Least Frequently Used
    42. #
    43. # Both LRU, LFU and volatile-ttl are implemented using approximated
    44. # randomized algorithms.
    45. #
    46. # Note: with any of the above policies, Redis will return an error on write
    47. #       operations, when there are no suitable keys for eviction.
    48. #
    49. #       At the date of writing these commands are: set setnx setex append
    50. #       incr decr rpush lpush rpushx lpushx linsert lset rpoplpush sadd
    51. #       sinter sinterstore sunion sunionstore sdiff sdiffstore zadd zincrby
    52. #       zunionstore zinterstore hset hsetnx hmset hincrby incrby decrby
    53. #       getset mset msetnx exec sort
    54. #
    55. # The default is:
    56. #
    57. # maxmemory-policy noeviction
    59. # LRU, LFU and minimal TTL algorithms are not precise algorithms but approximated
    60. # algorithms (in order to save memory), so you can tune it for speed or
    61. # accuracy. For default Redis will check five keys and pick the one that was
    62. # used less recently, you can change the sample size using the following
    63. # configuration directive.
    64. #
    65. # The default of 5 produces good enough results. 10 Approximates very closely
    66. # true LRU but costs more CPU. 3 is faster but not very accurate.
    67. #
    68. # maxmemory-samples 5

    LAZY FREEING

    1. ############################# LAZY FREEING ####################################
    3. # Redis has two primitives to delete keys. One is called DEL and is a blocking
    4. # deletion of the object. It means that the server stops processing new commands
    5. # in order to reclaim all the memory associated with an object in a synchronous
    6. # way. If the key deleted is associated with a small object, the time needed
    7. # in order to execute the DEL command is very small and comparable to most other
    8. # O(1) or O(log_N) commands in Redis. However if the key is associated with an
    9. # aggregated value containing millions of elements, the server can block for
    10. # a long time (even seconds) in order to complete the operation.
    11. #
    12. # For the above reasons Redis also offers non blocking deletion primitives
    13. # such as UNLINK (non blocking DEL) and the ASYNC option of FLUSHALL and
    14. # FLUSHDB commands, in order to reclaim memory in background. Those commands
    15. # are executed in constant time. Another thread will incrementally free the
    16. # object in the background as fast as possible.
    17. #
    18. # DEL, UNLINK and ASYNC option of FLUSHALL and FLUSHDB are user-controlled.
    19. # It's up to the design of the application to understand when it is a good
    20. # idea to use one or the other. However the Redis server sometimes has to
    21. # delete keys or flush the whole database as a side effect of other operations.
    22. # Specifically Redis deletes objects independently of a user call in the
    23. # following scenarios:
    24. #
    25. # 1) On eviction, because of the maxmemory and maxmemory policy configurations,
    26. #    in order to make room for new data, without going over the specified
    27. #    memory limit.
    28. # 2) Because of expire: when a key with an associated time to live (see the
    29. #    EXPIRE command) must be deleted from memory.
    30. # 3) Because of a side effect of a command that stores data on a key that may
    31. #    already exist. For example the RENAME command may delete the old key
    32. #    content when it is replaced with another one. Similarly SUNIONSTORE
    33. #    or SORT with STORE option may delete existing keys. The SET command
    34. #    itself removes any old content of the specified key in order to replace
    35. #    it with the specified string.
    36. # 4) During replication, when a slave performs a full resynchronization with
    37. #    its master, the content of the whole database is removed in order to
    38. #    load the RDB file just transfered.
    39. #
    40. # In all the above cases the default is to delete objects in a blocking way,
    41. # like if DEL was called. However you can configure each case specifically
    42. # in order to instead release memory in a non-blocking way like if UNLINK
    43. # was called, using the following configuration directives:
    45. lazyfree-lazy-eviction no
    46. lazyfree-lazy-expire no
    47. lazyfree-lazy-server-del no
    48. slave-lazy-flush no

    APPEND ONLY MODE

    1. ############################## APPEND ONLY MODE ###############################
    3. # By default Redis asynchronously dumps the dataset on disk. This mode is
    4. # good enough in many applications, but an issue with the Redis process or
    5. # a power outage may result into a few minutes of writes lost (depending on
    6. # the configured save points).
    7. #
    8. # The Append Only File is an alternative persistence mode that provides
    9. # much better durability. For instance using the default data fsync policy
    10. # (see later in the config file) Redis can lose just one second of writes in a
    11. # dramatic event like a server power outage, or a single write if something
    12. # wrong with the Redis process itself happens, but the operating system is
    13. # still running correctly.
    14. #
    15. # AOF and RDB persistence can be enabled at the same time without problems.
    16. # If the AOF is enabled on startup Redis will load the AOF, that is the file
    17. # with the better durability guarantees.
    18. #
    19. # Please check http://redis.io/topics/persistence for more information.
    21. appendonly no
    23. # The name of the append only file (default: "appendonly.aof")
    25. appendfilename "appendonly.aof"
    27. # The fsync() call tells the Operating System to actually write data on disk
    28. # instead of waiting for more data in the output buffer. Some OS will really flush
    29. # data on disk, some other OS will just try to do it ASAP.
    30. #
    31. # Redis supports three different modes:
    32. #
    33. # no: don't fsync, just let the OS flush the data when it wants. Faster.
    34. # always: fsync after every write to the append only log. Slow, Safest.
    35. # everysec: fsync only one time every second. Compromise.
    36. #
    37. # The default is "everysec", as that's usually the right compromise between
    38. # speed and data safety. It's up to you to understand if you can relax this to
    39. # "no" that will let the operating system flush the output buffer when
    40. # it wants, for better performances (but if you can live with the idea of
    41. # some data loss consider the default persistence mode that's snapshotting),
    42. # or on the contrary, use "always" that's very slow but a bit safer than
    43. # everysec.
    44. #
    45. # More details please check the following article:
    46. # http://antirez.com/post/redis-persistence-demystified.html
    47. #
    48. # If unsure, use "everysec".
    50. # appendfsync always
    51. appendfsync everysec
    52. # appendfsync no
    54. # When the AOF fsync policy is set to always or everysec, and a background
    55. # saving process (a background save or AOF log background rewriting) is
    56. # performing a lot of I/O against the disk, in some Linux configurations
    57. # Redis may block too long on the fsync() call. Note that there is no fix for
    58. # this currently, as even performing fsync in a different thread will block
    59. # our synchronous write(2) call.
    60. #
    61. # In order to mitigate this problem it's possible to use the following option
    62. # that will prevent fsync() from being called in the main process while a
    63. # BGSAVE or BGREWRITEAOF is in progress.
    64. #
    65. # This means that while another child is saving, the durability of Redis is
    66. # the same as "appendfsync none". In practical terms, this means that it is
    67. # possible to lose up to 30 seconds of log in the worst scenario (with the
    68. # default Linux settings).
    69. #
    70. # If you have latency problems turn this to "yes". Otherwise leave it as
    71. # "no" that is the safest pick from the point of view of durability.
    73. no-appendfsync-on-rewrite no
    75. # Automatic rewrite of the append only file.
    76. # Redis is able to automatically rewrite the log file implicitly calling
    77. # BGREWRITEAOF when the AOF log size grows by the specified percentage.
    78. #
    79. # This is how it works: Redis remembers the size of the AOF file after the
    80. # latest rewrite (if no rewrite has happened since the restart, the size of
    81. # the AOF at startup is used).
    82. #
    83. # This base size is compared to the current size. If the current size is
    84. # bigger than the specified percentage, the rewrite is triggered. Also
    85. # you need to specify a minimal size for the AOF file to be rewritten, this
    86. # is useful to avoid rewriting the AOF file even if the percentage increase
    87. # is reached but it is still pretty small.
    88. #
    89. # Specify a percentage of zero in order to disable the automatic AOF
    90. # rewrite feature.
    92. auto-aof-rewrite-percentage 100
    93. auto-aof-rewrite-min-size 64mb
    95. # An AOF file may be found to be truncated at the end during the Redis
    96. # startup process, when the AOF data gets loaded back into memory.
    97. # This may happen when the system where Redis is running
    98. # crashes, especially when an ext4 filesystem is mounted without the
    99. # data=ordered option (however this can't happen when Redis itself
    100. # crashes or aborts but the operating system still works correctly).
    101. #
    102. # Redis can either exit with an error when this happens, or load as much
    103. # data as possible (the default now) and start if the AOF file is found
    104. # to be truncated at the end. The following option controls this behavior.
    105. #
    106. # If aof-load-truncated is set to yes, a truncated AOF file is loaded and
    107. # the Redis server starts emitting a log to inform the user of the event.
    108. # Otherwise if the option is set to no, the server aborts with an error
    109. # and refuses to start. When the option is set to no, the user requires
    110. # to fix the AOF file using the "redis-check-aof" utility before to restart
    111. # the server.
    112. #
    113. # Note that if the AOF file will be found to be corrupted in the middle
    114. # the server will still exit with an error. This option only applies when
    115. # Redis will try to read more data from the AOF file but not enough bytes
    116. # will be found.
    117. aof-load-truncated yes
    119. # When rewriting the AOF file, Redis is able to use an RDB preamble in the
    120. # AOF file for faster rewrites and recoveries. When this option is turned
    121. # on the rewritten AOF file is composed of two different stanzas:
    122. #
    123. #   [RDB file][AOF tail]
    124. #
    125. # When loading Redis recognizes that the AOF file starts with the "REDIS"
    126. # string and loads the prefixed RDB file, and continues loading the AOF
    127. # tail.
    128. #
    129. # This is currently turned off by default in order to avoid the surprise
    130. # of a format change, but will at some point be used as the default.
    131. aof-use-rdb-preamble no

    LUA SCRIPTING

    1. ################################ LUA SCRIPTING  ###############################
    3. # Max execution time of a Lua script in milliseconds.
    4. #
    5. # If the maximum execution time is reached Redis will log that a script is
    6. # still in execution after the maximum allowed time and will start to
    7. # reply to queries with an error.
    8. #
    9. # When a long running script exceeds the maximum execution time only the
    10. # SCRIPT KILL and SHUTDOWN NOSAVE commands are available. The first can be
    11. # used to stop a script that did not yet called write commands. The second
    12. # is the only way to shut down the server in the case a write command was
    13. # already issued by the script but the user doesn't want to wait for the natural
    14. # termination of the script.
    15. #
    16. # Set it to 0 or a negative value for unlimited execution without warnings.
    17. lua-time-limit 5000

    REDIS CLUSTER

    1. ################################ REDIS CLUSTER  ###############################
    2. #
    3. # ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
    4. # WARNING EXPERIMENTAL: Redis Cluster is considered to be stable code, however
    5. # in order to mark it as "mature" we need to wait for a non trivial percentage
    6. # of users to deploy it in production.
    7. # ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
    8. #
    9. # Normal Redis instances can't be part of a Redis Cluster; only nodes that are
    10. # started as cluster nodes can. In order to start a Redis instance as a
    11. # cluster node enable the cluster support uncommenting the following:
    12. #
    13. # cluster-enabled yes
    15. # Every cluster node has a cluster configuration file. This file is not
    16. # intended to be edited by hand. It is created and updated by Redis nodes.
    17. # Every Redis Cluster node requires a different cluster configuration file.
    18. # Make sure that instances running in the same system do not have
    19. # overlapping cluster configuration file names.
    20. #
    21. # cluster-config-file nodes-6379.conf
    23. # Cluster node timeout is the amount of milliseconds a node must be unreachable
    24. # for it to be considered in failure state.
    25. # Most other internal time limits are multiple of the node timeout.
    26. #
    27. # cluster-node-timeout 15000
    29. # A slave of a failing master will avoid to start a failover if its data
    30. # looks too old.
    31. #
    32. # There is no simple way for a slave to actually have an exact measure of
    33. # its "data age", so the following two checks are performed:
    34. #
    35. # 1) If there are multiple slaves able to failover, they exchange messages
    36. #    in order to try to give an advantage to the slave with the best
    37. #    replication offset (more data from the master processed).
    38. #    Slaves will try to get their rank by offset, and apply to the start
    39. #    of the failover a delay proportional to their rank.
    40. #
    41. # 2) Every single slave computes the time of the last interaction with
    42. #    its master. This can be the last ping or command received (if the master
    43. #    is still in the "connected" state), or the time that elapsed since the
    44. #    disconnection with the master (if the replication link is currently down).
    45. #    If the last interaction is too old, the slave will not try to failover
    46. #    at all.
    47. #
    48. # The point "2" can be tuned by user. Specifically a slave will not perform
    49. # the failover if, since the last interaction with the master, the time
    50. # elapsed is greater than:
    51. #
    52. #   (node-timeout * slave-validity-factor) + repl-ping-slave-period
    53. #
    54. # So for example if node-timeout is 30 seconds, and the slave-validity-factor
    55. # is 10, and assuming a default repl-ping-slave-period of 10 seconds, the
    56. # slave will not try to failover if it was not able to talk with the master
    57. # for longer than 310 seconds.
    58. #
    59. # A large slave-validity-factor may allow slaves with too old data to failover
    60. # a master, while a too small value may prevent the cluster from being able to
    61. # elect a slave at all.
    62. #
    63. # For maximum availability, it is possible to set the slave-validity-factor
    64. # to a value of 0, which means, that slaves will always try to failover the
    65. # master regardless of the last time they interacted with the master.
    66. # (However they'll always try to apply a delay proportional to their
    67. # offset rank).
    68. #
    69. # Zero is the only value able to guarantee that when all the partitions heal
    70. # the cluster will always be able to continue.
    71. #
    72. # cluster-slave-validity-factor 10
    74. # Cluster slaves are able to migrate to orphaned masters, that are masters
    75. # that are left without working slaves. This improves the cluster ability
    76. # to resist to failures as otherwise an orphaned master can't be failed over
    77. # in case of failure if it has no working slaves.
    78. #
    79. # Slaves migrate to orphaned masters only if there are still at least a
    80. # given number of other working slaves for their old master. This number
    81. # is the "migration barrier". A migration barrier of 1 means that a slave
    82. # will migrate only if there is at least 1 other working slave for its master
    83. # and so forth. It usually reflects the number of slaves you want for every
    84. # master in your cluster.
    85. #
    86. # Default is 1 (slaves migrate only if their masters remain with at least
    87. # one slave). To disable migration just set it to a very large value.
    88. # A value of 0 can be set but is useful only for debugging and dangerous
    89. # in production.
    90. #
    91. # cluster-migration-barrier 1
    93. # By default Redis Cluster nodes stop accepting queries if they detect there
    94. # is at least an hash slot uncovered (no available node is serving it).
    95. # This way if the cluster is partially down (for example a range of hash slots
    96. # are no longer covered) all the cluster becomes, eventually, unavailable.
    97. # It automatically returns available as soon as all the slots are covered again.
    98. #
    99. # However sometimes you want the subset of the cluster which is working,
    100. # to continue to accept queries for the part of the key space that is still
    101. # covered. In order to do so, just set the cluster-require-full-coverage
    102. # option to no.
    103. #
    104. # cluster-require-full-coverage yes
    106. # In order to setup your cluster make sure to read the documentation
    107. # available at http://redis.io web site.

    CLUSTER DOCKER/NAT support

    1. ########################## CLUSTER DOCKER/NAT support  ########################
    3. # In certain deployments, Redis Cluster nodes address discovery fails, because
    4. # addresses are NAT-ted or because ports are forwarded (the typical case is
    5. # Docker and other containers).
    6. #
    7. # In order to make Redis Cluster working in such environments, a static
    8. # configuration where each node knows its public address is needed. The
    9. # following two options are used for this scope, and are:
    10. #
    11. # * cluster-announce-ip
    12. # * cluster-announce-port
    13. # * cluster-announce-bus-port
    14. #
    15. # Each instruct the node about its address, client port, and cluster message
    16. # bus port. The information is then published in the header of the bus packets
    17. # so that other nodes will be able to correctly map the address of the node
    18. # publishing the information.
    19. #
    20. # If the above options are not used, the normal Redis Cluster auto-detection
    21. # will be used instead.
    22. #
    23. # Note that when remapped, the bus port may not be at the fixed offset of
    24. # clients port + 10000, so you can specify any port and bus-port depending
    25. # on how they get remapped. If the bus-port is not set, a fixed offset of
    26. # 10000 will be used as usually.
    27. #
    28. # Example:
    29. #
    30. # cluster-announce-ip 10.1.1.5
    31. # cluster-announce-port 6379
    32. # cluster-announce-bus-port 6380

    SLOW LOG

    1. ################################## SLOW LOG ###################################
    3. # The Redis Slow Log is a system to log queries that exceeded a specified
    4. # execution time. The execution time does not include the I/O operations
    5. # like talking with the client, sending the reply and so forth,
    6. # but just the time needed to actually execute the command (this is the only
    7. # stage of command execution where the thread is blocked and can not serve
    8. # other requests in the meantime).
    9. #
    10. # You can configure the slow log with two parameters: one tells Redis
    11. # what is the execution time, in microseconds, to exceed in order for the
    12. # command to get logged, and the other parameter is the length of the
    13. # slow log. When a new command is logged the oldest one is removed from the
    14. # queue of logged commands.
    16. # The following time is expressed in microseconds, so 1000000 is equivalent
    17. # to one second. Note that a negative number disables the slow log, while
    18. # a value of zero forces the logging of every command.
    19. slowlog-log-slower-than 10000
    21. # There is no limit to this length. Just be aware that it will consume memory.
    22. # You can reclaim memory used by the slow log with SLOWLOG RESET.
    23. slowlog-max-len 128

    LATENCY MONITOR

    1. ################################ LATENCY MONITOR ##############################
    3. # The Redis latency monitoring subsystem samples different operations
    4. # at runtime in order to collect data related to possible sources of
    5. # latency of a Redis instance.
    6. #
    7. # Via the LATENCY command this information is available to the user that can
    8. # print graphs and obtain reports.
    9. #
    10. # The system only logs operations that were performed in a time equal or
    11. # greater than the amount of milliseconds specified via the
    12. # latency-monitor-threshold configuration directive. When its value is set
    13. # to zero, the latency monitor is turned off.
    14. #
    15. # By default latency monitoring is disabled since it is mostly not needed
    16. # if you don't have latency issues, and collecting data has a performance
    17. # impact, that while very small, can be measured under big load. Latency
    18. # monitoring can easily be enabled at runtime using the command
    19. # "CONFIG SET latency-monitor-threshold <milliseconds>" if needed.
    20. latency-monitor-threshold 0

    EVENT NOTIFICATION

    1. ############################# EVENT NOTIFICATION ##############################
    3. # Redis can notify Pub/Sub clients about events happening in the key space.
    4. # This feature is documented at http://redis.io/topics/notifications
    5. #
    6. # For instance if keyspace events notification is enabled, and a client
    7. # performs a DEL operation on key "foo" stored in the Database 0, two
    8. # messages will be published via Pub/Sub:
    9. #
    10. # PUBLISH __keyspace@0__:foo del
    11. # PUBLISH __keyevent@0__:del foo
    12. #
    13. # It is possible to select the events that Redis will notify among a set
    14. # of classes. Every class is identified by a single character:
    15. #
    16. #  K     Keyspace events, published with __keyspace@<db>__ prefix.
    17. #  E     Keyevent events, published with __keyevent@<db>__ prefix.
    18. #  g     Generic commands (non-type specific) like DEL, EXPIRE, RENAME, ...
    19. #  $     String commands
    20. #  l     List commands
    21. #  s     Set commands
    22. #  h     Hash commands
    23. #  z     Sorted set commands
    24. #  x     Expired events (events generated every time a key expires)
    25. #  e     Evicted events (events generated when a key is evicted for maxmemory)
    26. #  A     Alias for g$lshzxe, so that the "AKE" string means all the events.
    27. #
    28. #  The "notify-keyspace-events" takes as argument a string that is composed
    29. #  of zero or multiple characters. The empty string means that notifications
    30. #  are disabled.
    31. #
    32. #  Example: to enable list and generic events, from the point of view of the
    33. #           event name, use:
    34. #
    35. #  notify-keyspace-events Elg
    36. #
    37. #  Example 2: to get the stream of the expired keys subscribing to channel
    38. #             name __keyevent@0__:expired use:
    39. #
    40. #  notify-keyspace-events Ex
    41. #
    42. #  By default all notifications are disabled because most users don't need
    43. #  this feature and the feature has some overhead. Note that if you don't
    44. #  specify at least one of K or E, no events will be delivered.
    45. notify-keyspace-events ""

    ADVANCED CONFIG

    1. ############################### ADVANCED CONFIG ###############################
    3. # Hashes are encoded using a memory efficient data structure when they have a
    4. # small number of entries, and the biggest entry does not exceed a given
    5. # threshold. These thresholds can be configured using the following directives.
    6. hash-max-ziplist-entries 512
    7. hash-max-ziplist-value 64
    9. # Lists are also encoded in a special way to save a lot of space.
    10. # The number of entries allowed per internal list node can be specified
    11. # as a fixed maximum size or a maximum number of elements.
    12. # For a fixed maximum size, use -5 through -1, meaning:
    13. # -5: max size: 64 Kb  <-- not recommended for normal workloads
    14. # -4: max size: 32 Kb  <-- not recommended
    15. # -3: max size: 16 Kb  <-- probably not recommended
    16. # -2: max size: 8 Kb   <-- good
    17. # -1: max size: 4 Kb   <-- good
    18. # Positive numbers mean store up to _exactly_ that number of elements
    19. # per list node.
    20. # The highest performing option is usually -2 (8 Kb size) or -1 (4 Kb size),
    21. # but if your use case is unique, adjust the settings as necessary.
    22. list-max-ziplist-size -2
    24. # Lists may also be compressed.
    25. # Compress depth is the number of quicklist ziplist nodes from *each* side of
    26. # the list to *exclude* from compression.  The head and tail of the list
    27. # are always uncompressed for fast push/pop operations.  Settings are:
    28. # 0: disable all list compression
    29. # 1: depth 1 means "don't start compressing until after 1 node into the list,
    30. #    going from either the head or tail"
    31. #    So: [head]->node->node->...->node->[tail]
    32. #    [head], [tail] will always be uncompressed; inner nodes will compress.
    33. # 2: [head]->[next]->node->node->...->node->[prev]->[tail]
    34. #    2 here means: don't compress head or head->next or tail->prev or tail,
    35. #    but compress all nodes between them.
    36. # 3: [head]->[next]->[next]->node->node->...->node->[prev]->[prev]->[tail]
    37. # etc.
    38. list-compress-depth 0
    40. # Sets have a special encoding in just one case: when a set is composed
    41. # of just strings that happen to be integers in radix 10 in the range
    42. # of 64 bit signed integers.
    43. # The following configuration setting sets the limit in the size of the
    44. # set in order to use this special memory saving encoding.
    45. set-max-intset-entries 512
    47. # Similarly to hashes and lists, sorted sets are also specially encoded in
    48. # order to save a lot of space. This encoding is only used when the length and
    49. # elements of a sorted set are below the following limits:
    50. zset-max-ziplist-entries 128
    51. zset-max-ziplist-value 64
    53. # HyperLogLog sparse representation bytes limit. The limit includes the
    54. # 16 bytes header. When an HyperLogLog using the sparse representation crosses
    55. # this limit, it is converted into the dense representation.
    56. #
    57. # A value greater than 16000 is totally useless, since at that point the
    58. # dense representation is more memory efficient.
    59. #
    60. # The suggested value is ~ 3000 in order to have the benefits of
    61. # the space efficient encoding without slowing down too much PFADD,
    62. # which is O(N) with the sparse encoding. The value can be raised to
    63. # ~ 10000 when CPU is not a concern, but space is, and the data set is
    64. # composed of many HyperLogLogs with cardinality in the 0 - 15000 range.
    65. hll-sparse-max-bytes 3000
    67. # Active rehashing uses 1 millisecond every 100 milliseconds of CPU time in
    68. # order to help rehashing the main Redis hash table (the one mapping top-level
    69. # keys to values). The hash table implementation Redis uses (see dict.c)
    70. # performs a lazy rehashing: the more operation you run into a hash table
    71. # that is rehashing, the more rehashing "steps" are performed, so if the
    72. # server is idle the rehashing is never complete and some more memory is used
    73. # by the hash table.
    74. #
    75. # The default is to use this millisecond 10 times every second in order to
    76. # actively rehash the main dictionaries, freeing memory when possible.
    77. #
    78. # If unsure:
    79. # use "activerehashing no" if you have hard latency requirements and it is
    80. # not a good thing in your environment that Redis can reply from time to time
    81. # to queries with 2 milliseconds delay.
    82. #
    83. # use "activerehashing yes" if you don't have such hard requirements but
    84. # want to free memory asap when possible.
    85. activerehashing yes
    87. # The client output buffer limits can be used to force disconnection of clients
    88. # that are not reading data from the server fast enough for some reason (a
    89. # common reason is that a Pub/Sub client can't consume messages as fast as the
    90. # publisher can produce them).
    91. #
    92. # The limit can be set differently for the three different classes of clients:
    93. #
    94. # normal -> normal clients including MONITOR clients
    95. # slave  -> slave clients
    96. # pubsub -> clients subscribed to at least one pubsub channel or pattern
    97. #
    98. # The syntax of every client-output-buffer-limit directive is the following:
    99. #
    100. # client-output-buffer-limit <class> <hard limit> <soft limit> <soft seconds>
    101. #
    102. # A client is immediately disconnected once the hard limit is reached, or if
    103. # the soft limit is reached and remains reached for the specified number of
    104. # seconds (continuously).
    105. # So for instance if the hard limit is 32 megabytes and the soft limit is
    106. # 16 megabytes / 10 seconds, the client will get disconnected immediately
    107. # if the size of the output buffers reach 32 megabytes, but will also get
    108. # disconnected if the client reaches 16 megabytes and continuously overcomes
    109. # the limit for 10 seconds.
    110. #
    111. # By default normal clients are not limited because they don't receive data
    112. # without asking (in a push way), but just after a request, so only
    113. # asynchronous clients may create a scenario where data is requested faster
    114. # than it can read.
    115. #
    116. # Instead there is a default limit for pubsub and slave clients, since
    117. # subscribers and slaves receive data in a push fashion.
    118. #
    119. # Both the hard or the soft limit can be disabled by setting them to zero.
    120. client-output-buffer-limit normal 0 0 0
    121. client-output-buffer-limit slave 256mb 64mb 60
    122. client-output-buffer-limit pubsub 32mb 8mb 60
    124. # Client query buffers accumulate new commands. They are limited to a fixed
    125. # amount by default in order to avoid that a protocol desynchronization (for
    126. # instance due to a bug in the client) will lead to unbound memory usage in
    127. # the query buffer. However you can configure it here if you have very special
    128. # needs, such us huge multi/exec requests or alike.
    129. #
    130. # client-query-buffer-limit 1gb
    132. # In the Redis protocol, bulk requests, that are, elements representing single
    133. # strings, are normally limited ot 512 mb. However you can change this limit
    134. # here.
    135. #
    136. # proto-max-bulk-len 512mb
    138. # Redis calls an internal function to perform many background tasks, like
    139. # closing connections of clients in timeout, purging expired keys that are
    140. # never requested, and so forth.
    141. #
    142. # Not all tasks are performed with the same frequency, but Redis checks for
    143. # tasks to perform according to the specified "hz" value.
    144. #
    145. # By default "hz" is set to 10. Raising the value will use more CPU when
    146. # Redis is idle, but at the same time will make Redis more responsive when
    147. # there are many keys expiring at the same time, and timeouts may be
    148. # handled with more precision.
    149. #
    150. # The range is between 1 and 500, however a value over 100 is usually not
    151. # a good idea. Most users should use the default of 10 and raise this up to
    152. # 100 only in environments where very low latency is required.
    153. hz 10
    155. # When a child rewrites the AOF file, if the following option is enabled
    156. # the file will be fsync-ed every 32 MB of data generated. This is useful
    157. # in order to commit the file to the disk more incrementally and avoid
    158. # big latency spikes.
    159. aof-rewrite-incremental-fsync yes
    161. # Redis LFU eviction (see maxmemory setting) can be tuned. However it is a good
    162. # idea to start with the default settings and only change them after investigating
    163. # how to improve the performances and how the keys LFU change over time, which
    164. # is possible to inspect via the OBJECT FREQ command.
    165. #
    166. # There are two tunable parameters in the Redis LFU implementation: the
    167. # counter logarithm factor and the counter decay time. It is important to
    168. # understand what the two parameters mean before changing them.
    169. #
    170. # The LFU counter is just 8 bits per key, it's maximum value is 255, so Redis
    171. # uses a probabilistic increment with logarithmic behavior. Given the value
    172. # of the old counter, when a key is accessed, the counter is incremented in
    173. # this way:
    174. #
    175. # 1. A random number R between 0 and 1 is extracted.
    176. # 2. A probability P is calculated as 1/(old_value*lfu_log_factor+1).
    177. # 3. The counter is incremented only if R < P.
    178. #
    179. # The default lfu-log-factor is 10. This is a table of how the frequency
    180. # counter changes with a different number of accesses with different
    181. # logarithmic factors:
    182. #
    183. # +--------+------------+------------+------------+------------+------------+
    184. # | factor | 100 hits   | 1000 hits  | 100K hits  | 1M hits    | 10M hits   |
    185. # +--------+------------+------------+------------+------------+------------+
    186. # | 0      | 104        | 255        | 255        | 255        | 255        |
    187. # +--------+------------+------------+------------+------------+------------+
    188. # | 1      | 18         | 49         | 255        | 255        | 255        |
    189. # +--------+------------+------------+------------+------------+------------+
    190. # | 10     | 10         | 18         | 142        | 255        | 255        |
    191. # +--------+------------+------------+------------+------------+------------+
    192. # | 100    | 8          | 11         | 49         | 143        | 255        |
    193. # +--------+------------+------------+------------+------------+------------+
    194. #
    195. # NOTE: The above table was obtained by running the following commands:
    196. #
    197. #   redis-benchmark -n 1000000 incr foo
    198. #   redis-cli object freq foo
    199. #
    200. # NOTE 2: The counter initial value is 5 in order to give new objects a chance
    201. # to accumulate hits.
    202. #
    203. # The counter decay time is the time, in minutes, that must elapse in order
    204. # for the key counter to be divided by two (or decremented if it has a value
    205. # less <= 10).
    206. #
    207. # The default value for the lfu-decay-time is 1. A Special value of 0 means to
    208. # decay the counter every time it happens to be scanned.
    209. #
    210. # lfu-log-factor 10
    211. # lfu-decay-time 1

    ACTIVE DEFRAGMENTATION

    1. ########################### ACTIVE DEFRAGMENTATION #######################
    2. #
    3. # WARNING THIS FEATURE IS EXPERIMENTAL. However it was stress tested
    4. # even in production and manually tested by multiple engineers for some
    5. # time.
    6. #
    7. # What is active defragmentation?
    8. # -------------------------------
    9. #
    10. # Active (online) defragmentation allows a Redis server to compact the
    11. # spaces left between small allocations and deallocations of data in memory,
    12. # thus allowing to reclaim back memory.
    13. #
    14. # Fragmentation is a natural process that happens with every allocator (but
    15. # less so with Jemalloc, fortunately) and certain workloads. Normally a server
    16. # restart is needed in order to lower the fragmentation, or at least to flush
    17. # away all the data and create it again. However thanks to this feature
    18. # implemented by Oran Agra for Redis 4.0 this process can happen at runtime
    19. # in an "hot" way, while the server is running.
    20. #
    21. # Basically when the fragmentation is over a certain level (see the
    22. # configuration options below) Redis will start to create new copies of the
    23. # values in contiguous memory regions by exploiting certain specific Jemalloc
    24. # features (in order to understand if an allocation is causing fragmentation
    25. # and to allocate it in a better place), and at the same time, will release the
    26. # old copies of the data. This process, repeated incrementally for all the keys
    27. # will cause the fragmentation to drop back to normal values.
    28. #
    29. # Important things to understand:
    30. #
    31. # 1. This feature is disabled by default, and only works if you compiled Redis
    32. #    to use the copy of Jemalloc we ship with the source code of Redis.
    33. #    This is the default with Linux builds.
    34. #
    35. # 2. You never need to enable this feature if you don't have fragmentation
    36. #    issues.
    37. #
    38. # 3. Once you experience fragmentation, you can enable this feature when
    39. #    needed with the command "CONFIG SET activedefrag yes".
    40. #
    41. # The configuration parameters are able to fine tune the behavior of the
    42. # defragmentation process. If you are not sure about what they mean it is
    43. # a good idea to leave the defaults untouched.
    45. # Enabled active defragmentation
    46. # activedefrag yes
    48. # Minimum amount of fragmentation waste to start active defrag
    49. # active-defrag-ignore-bytes 100mb
    51. # Minimum percentage of fragmentation to start active defrag
    52. # active-defrag-threshold-lower 10
    54. # Maximum percentage of fragmentation at which we use maximum effort
    55. # active-defrag-threshold-upper 100
    57. # Minimal effort for defrag in CPU percentage
    58. # active-defrag-cycle-min 25
    60. # Maximal effort for defrag in CPU percentage
    61. # active-defrag-cycle-max 75