# Valkey configuration file example.

#

# Note that in order to read the configuration file, the server must be

# started with the file path as first argument:

#

# ./valkey-server /path/to/valkey.conf

# Note on units: when memory size is needed, it is possible to specify

# it in the usual form of 1k 5GB 4M and so forth:

#

# 1k => 1000 bytes

# 1kb => 1024 bytes

# 1m => 1000000 bytes

# 1mb => 1024*1024 bytes

# 1g => 1000000000 bytes

# 1gb => 1024*1024*1024 bytes

#

# units are case insensitive so 1GB 1Gb 1gB are all the same.

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

# Include one or more other config files here.  This is useful if you

# have a standard template that goes to all servers but also need

# to customize a few per-server settings.  Include files can include

# other files, so use this wisely.

#

# Note that option "include" won't be rewritten by command "CONFIG REWRITE"

# from admin or Sentinel. Since the server always uses the last processed

# line as value of a configuration directive, you'd better put includes

# at the beginning of this file to avoid overwriting config change at runtime.

#

# If instead you are interested in using includes to override configuration

# options, it is better to use include as the last line.

#

# Included paths may contain wildcards. All files matching the wildcards will

# be included in alphabetical order.

# Note that if an include path contains a wildcards but no files match it when

# the server is started, the include statement will be ignored and no error will

# be emitted.  It is safe, therefore, to include wildcard files from empty

# directories.

#

# include /path/to/local.conf

# include /path/to/other.conf

# include /path/to/fragments/*.conf

#

################################## MODULES #####################################

# Load modules at startup. If the server is not able to load modules

# it will abort. It is possible to use multiple loadmodule directives.

#

# loadmodule /path/to/my_module.so

# loadmodule /path/to/other_module.so

# loadmodule /path/to/args_module.so [arg [arg ...]]

################################## NETWORK #####################################

# By default, if no "bind" configuration directive is specified, the server listens

# for connections from all available network interfaces on the host machine.

# It is possible to listen to just one or multiple selected interfaces using

# the "bind" configuration directive, followed by one or more IP addresses.

# Each address can be prefixed by "-", which means that the server will not fail to

# start if the address is not available. Being not available only refers to

# addresses that does not correspond to any network interface. Addresses that

# are already in use will always fail, and unsupported protocols will always be

# silently skipped.

#

# Examples:

#

# bind 192.168.1.100 10.0.0.1     # listens on two specific IPv4 addresses

# bind 127.0.0.1 ::1              # listens on loopback IPv4 and IPv6

# bind * -::*                     # like the default, all available interfaces

#

# ~~~ WARNING ~~~ If the computer running the server is directly exposed to the

# internet, binding to all the interfaces is dangerous and will expose the

# instance to everybody on the internet. So by default we uncomment the

# following bind directive, that will force the server to listen only on the

# IPv4 and IPv6 (if available) loopback interface addresses (this means the server

# will only be able to accept client connections from the same host that it is

# running on).

#

# IF YOU ARE SURE YOU WANT YOUR INSTANCE TO LISTEN TO ALL THE INTERFACES

# COMMENT OUT THE FOLLOWING LINE.

#

# You will also need to set a password unless you explicitly disable protected

# mode.

# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

bind 0.0.0.0 -::1

# By default, outgoing connections (from replica to primary, from Sentinel to

# instances, cluster bus, etc.) are not bound to a specific local address. In

# most cases, this means the operating system will handle that based on routing

# and the interface through which the connection goes out.

#

# Using bind-source-addr it is possible to configure a specific address to bind

# to, which may also affect how the connection gets routed.

#

# Example:

#

# bind-source-addr 10.0.0.1

# Protected mode is a layer of security protection, in order to avoid that

# the server instances left open on the internet are accessed and exploited.

#

# When protected mode is on and the default user has no password, the server

# only accepts local connections from the IPv4 address (127.0.0.1), IPv6 address

# (::1) or Unix domain sockets.

#

# By default protected mode is enabled. You should disable it only if

# you are sure you want clients from other hosts to connect to the server

# even if no authentication is configured.

protected-mode no

# The server uses default hardened security configuration directives to reduce the

# attack surface on innocent users. Therefore, several sensitive configuration

# directives are immutable, and some potentially-dangerous commands are blocked.

#

# Configuration directives that control files that the server writes to (e.g., 'dir'

# and 'dbfilename') and that aren't usually modified during runtime

# are protected by making them immutable.

#

# Commands that can increase the attack surface of the server and that aren't usually

# called by users are blocked by default.

#

# These can be exposed to either all connections or just local ones by setting

# each of the configs listed below to either of these values:

#

# no    - Block for any connection (remain immutable)

# yes   - Allow for any connection (no protection)

# local - Allow only for local connections. Ones originating from the

#         IPv4 address (127.0.0.1), IPv6 address (::1) or Unix domain sockets.

#

# enable-protected-configs no

# enable-debug-command no

# enable-module-command no

# Accept connections on the specified port, default is 6379 (IANA #815344).

# If port 0 is specified the server will not listen on a TCP socket.

port 6379

# TCP listen() backlog.

#

# In high requests-per-second environments you need a high backlog in order

# to avoid slow clients connection issues. Note that the Linux kernel

# will silently truncate it to the value of /proc/sys/net/core/somaxconn so

# make sure to raise both the value of somaxconn and tcp_max_syn_backlog

# in order to get the desired effect.

tcp-backlog 511

# Multipath TCP (MPTCP)

#

# MPTCP splits a single TCP connection into subflows over multiple interfaces or paths.

# It enables bandwidth aggregation, failover, and improved reliability.

# When set to 'yes', clients will be able to use MPTCP if requested. When not

# requested, regular TCP can be used like before.

# Note: MPTCP is supported in the mainline Linux kernel starting from version 5.6.

#

# mptcp yes

# Unix socket.

#

# Specify the path for the Unix socket that will be used to listen for

# incoming connections. There is no default, so the server will not listen

# on a unix socket when not specified.

#

# unixsocket /run/valkey.sock

# unixsocketgroup wheel

# unixsocketperm 700

# Close the connection after a client is idle for N seconds (0 to disable)

timeout 0

# TCP keepalive.

#

# If non-zero, use SO_KEEPALIVE to send TCP ACKs to clients in absence

# of communication. This is useful for two reasons:

#

# 1) Detect dead peers.

# 2) Force network equipment in the middle to consider the connection to be

#    alive.

#

# On Linux, the specified value (in seconds) is the period used to send ACKs.

# Note that to close the connection the double of the time is needed.

# On other kernels the period depends on the kernel configuration.

tcp-keepalive 300

# Apply OS-specific mechanism to mark the listening socket with the specified

# ID, to support advanced routing and filtering capabilities.

#

# On Linux, the ID represents a connection mark.

# On FreeBSD, the ID represents a socket cookie ID.

# On OpenBSD, the ID represents a route table ID.

#

# The default value is 0, which implies no marking is required.

# socket-mark-id 0

################################# TLS/SSL #####################################

# By default, TLS/SSL is disabled. To enable it, the "tls-port" configuration

# directive can be used to define TLS-listening ports. To enable TLS on the

# default port, use:

#

# port 0

# tls-port 6379

# Configure a X.509 certificate and private key to use for authenticating the

# server to connected clients, primaries or cluster peers.  These files should be

# PEM formatted.

#

# tls-cert-file valkey.crt

# tls-key-file valkey.key

#

# If the key file is encrypted using a passphrase, it can be included here

# as well.

#

# tls-key-file-pass secret

# Normally the server uses the same certificate for both server functions (accepting

# connections) and client functions (replicating from a primary, establishing

# cluster bus connections, etc.).

#

# Sometimes certificates are issued with attributes that designate them as

# client-only or server-only certificates. In that case it may be desired to use

# different certificates for incoming (server) and outgoing (client)

# connections. To do that, use the following directives:

#

# tls-client-cert-file client.crt

# tls-client-key-file client.key

#

# If the key file is encrypted using a passphrase, it can be included here

# as well.

#

# tls-client-key-file-pass secret

# Configure a DH parameters file to enable Diffie-Hellman (DH) key exchange,

# required by older versions of OpenSSL (<3.0). Newer versions do not require

# this configuration and recommend against it.

#

# tls-dh-params-file valkey.dh

# Configure a CA certificate(s) bundle or directory to authenticate TLS/SSL

# clients and peers. The server requires an explicit configuration of at least one

# of these, and will not implicitly use the system wide configuration.

#

# tls-ca-cert-file ca.crt

# tls-ca-cert-dir /etc/ssl/certs

# By default, clients (including replica servers) on a TLS port are required

# to authenticate using valid client side certificates.

#

# If "no" is specified, client certificates are not required and not accepted.

# If "optional" is specified, client certificates are accepted and must be

# valid if provided, but are not required.

#

# tls-auth-clients no

# tls-auth-clients optional

# Automatically authenticate TLS clients as Valkey users based on their

# certificates.

#

# If set to a field like "CN", the server will extract the corresponding field

# from the client's TLS certificate and attempt to find a Valkey user with the

# same name. If a matching user is found, the client is automatically

# authenticated as that user during the TLS handshake. If no matching user is

# found, the client is connected as the unauthenticated default user. Set to

# "off" to disable automatic user authentication via certificate fields.

#

# Supported values: CN, off. Default: off.

#

# tls-auth-clients-user CN

# By default, a replica does not attempt to establish a TLS connection

# with its primary.

#

# Use the following directive to enable TLS on replication links.

#

# tls-replication yes

# By default, the cluster bus uses a plain TCP connection. To enable

# TLS for the bus protocol, use the following directive:

#

# tls-cluster yes

# By default, only TLSv1.2 and TLSv1.3 are enabled and it is highly recommended

# that older formally deprecated versions are kept disabled to reduce the attack surface.

# You can explicitly specify TLS versions to support.

# Allowed values are case insensitive and include "TLSv1", "TLSv1.1", "TLSv1.2",

# "TLSv1.3" (OpenSSL >= 1.1.1) or any combination.

# To enable only TLSv1.2 and TLSv1.3, use:

#

# tls-protocols "TLSv1.2 TLSv1.3"

# Configure allowed ciphers.  See the ciphers(1ssl) manpage for more information

# about the syntax of this string.

#

# Note: this configuration applies only to <= TLSv1.2.

#

# tls-ciphers DEFAULT:!MEDIUM

# Configure allowed TLSv1.3 ciphersuites.  See the ciphers(1ssl) manpage for more

# information about the syntax of this string, and specifically for TLSv1.3

# ciphersuites.

#

# tls-ciphersuites TLS_CHACHA20_POLY1305_SHA256

# When choosing a cipher, use the server's preference instead of the client

# preference. By default, the server follows the client's preference.

#

# tls-prefer-server-ciphers yes

# By default, TLS session caching is enabled to allow faster and less expensive

# reconnections by clients that support it. Use the following directive to disable

# caching.

#

# tls-session-caching no

# Change the default number of TLS sessions cached. A zero value sets the cache

# to unlimited size. The default size is 20480.

#

# tls-session-cache-size 5000

# Change the default timeout of cached TLS sessions. The default timeout is 300

# seconds.

#

# tls-session-cache-timeout 60

################################### RDMA ######################################

# Valkey Over RDMA is experimental, it may be changed or be removed in any minor or major version.

# By default, RDMA is disabled. To enable it, the "rdma-port" configuration

# directive can be used to define RDMA-listening ports.

#

# rdma-port 6379

# rdma-bind 192.168.1.100

# The RDMA receive transfer buffer is 1M by default. It can be set between 64K and 16M.

# Note that page size aligned size is preferred.

#

# rdma-rx-size 1048576

# The RDMA completion queue will use the completion vector to signal completion events

# via hardware interrupts. A large number of hardware interrupts can affect CPU performance.

# It is possible to tune the performance using rdma-completion-vector.

#

# Example 1. a) Pin hardware interrupt vectors [0, 3] to CPU [0, 3].

#            b) Set CPU affinity for valkey to CPU [4, X].

#            c) Any valkey server uses a random RDMA completion vector [-1].

# All valkey servers will not affect each other and will be isolated from kernel interrupts.

#

#   SYS    SYS    SYS    SYS  VALKEY VALKEY     VALKEY

#    |      |      |      |      |      |          |

#  CPU0   CPU1   CPU2   CPU3   CPU4   CPU5   ... CPUX

#    |      |      |      |

#  INTR0  INTR1  INTR2  INTR3

#

# Example 2. a) 1:1 pin hardware interrupt vectors [0, X] to CPU [0, X].

#            b) Set CPU affinity for valkey [M] to CPU [M].

#            c) Valkey server [M] uses RDMA completion vector [M].

# A single CPU [M] handles hardware interrupts, the RDMA completion vector [M],

# and the valkey server [M] within its context only.

# This avoids overhead and function calls across multiple CPUs, fully isolating

# each valkey server from one another.

#

# VALKEY VALKEY VALKEY VALKEY VALKEY VALKEY     VALKEY

#    |      |      |      |      |      |          |

#  CPU0   CPU1   CPU2   CPU3   CPU4   CPU5  ...  CPUX

#    |      |      |      |      |      |          |

#  INTR0  INTR1  INTR2  INTR3  INTR4  INTR5      INTRX

#

# Use 0 and positive numbers to specify the RDMA completion vector, or specify -1 to allow

# the server to use a random vector for a new connection. The default vector is -1.

#

# rdma-completion-vector 0

################################# GENERAL #####################################

# By default the server does not run as a daemon. Use 'yes' if you need it.

# Note that the server will write a pid file in /var/run/valkey.pid when daemonized.

# When the server is supervised by upstart or systemd, this parameter has no impact.

daemonize no

# If you run the server from upstart or systemd, the server can interact with your

# supervision tree. Options:

#   supervised no      - no supervision interaction

#   supervised upstart - signal upstart by putting the server into SIGSTOP mode

#                        requires "expect stop" in your upstart job config

#   supervised systemd - signal systemd by writing READY=1 to $NOTIFY_SOCKET

#                        on startup, and updating the server status on a regular

#                        basis.

#   supervised auto    - detect upstart or systemd method based on

#                        UPSTART_JOB or NOTIFY_SOCKET environment variables

# Note: these supervision methods only signal "process is ready."

#       They do not enable continuous pings back to your supervisor.

#

# The default is "no". To run under upstart/systemd, you can simply uncomment

# the line below:

#

# supervised auto

# If a pid file is specified, the server writes it where specified at startup

# and removes it at exit.

#

# When the server runs non daemonized, no pid file is created if none is

# specified in the configuration. When the server is daemonized, the pid file

# is used even if not specified, defaulting to "/var/run/valkey.pid".

#

# Creating a pid file is best effort: if the server is not able to create it

# nothing bad happens, the server will start and run normally.

#

# Note that on modern Linux systems "/run/valkey.pid" is more conforming

# and should be used instead.

pidfile /var/run/valkey_6379.pid

# Specify the server verbosity level.

# This can be one of:

# debug (a lot of information, useful for development/testing)

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

# notice (moderately verbose, what you want in production probably)

# warning (only very important / critical messages are logged)

# nothing (nothing is logged)

loglevel notice

# Specify the logging format.

# This can be one of:

#

# - legacy: the default, traditional log format

# - logfmt: a structured log format; see https://www.brandur.org/logfmt

#

# log-format legacy

# Specify the timestamp format used in logs using 'log-timestamp-format'.

#

# - legacy: default format

# - iso8601: ISO 8601 extended date and time with time zone, on the form

#   yyyy-mm-ddThh:mm:ss.sss±hh:mm

# - milliseconds: milliseconds since the epoch

#

# log-timestamp-format legacy

# Specify the log file name. Also the empty string can be used to force

# the server to log on the standard output. Note that if you use standard

# output for logging but daemonize, logs will be sent to /dev/null

logfile ""

# To enable logging to the system logger, just set 'syslog-enabled' to yes,

# and optionally update the other syslog parameters to suit your needs.

# syslog-enabled no

# Specify the syslog identity.

# syslog-ident valkey

# Specify the syslog facility. Must be USER or between LOCAL0-LOCAL7.

# syslog-facility local0

# To disable the built in crash log, which will possibly produce cleaner core

# dumps when they are needed, uncomment the following:

#

# crash-log-enabled no

# To disable the fast memory check that's run as part of the crash log, which

# will possibly let the server terminate sooner, uncomment the following:

#

# crash-memcheck-enabled no

# Set the number of databases. The default database is DB 0, you can select

# a different one on a per-connection basis using SELECT <dbid> where

# dbid is a number between 0 and 'databases'-1

# Note: This setting is ignored in cluster mode. Use `cluster-databases` instead.

databases 16

# By default the server shows an ASCII art logo only when started to log to the

# standard output and if the standard output is a TTY and syslog logging is

# disabled. Basically this means that normally a logo is displayed only in

# interactive sessions.

#

# However it is possible to force the pre-4.0 behavior and always show a

# ASCII art logo in startup logs by setting the following option to yes.

always-show-logo no

# User data, including keys, values, client names, and ACL usernames, can be

# logged as part of assertions and other error cases. To prevent sensitive user

# information, such as PII, from being recorded in the server log file, this

# user data is hidden from the log by default. If you need to log user data for

# debugging or troubleshooting purposes, you can disable this feature by

# changing the config value to no.

hide-user-data-from-log yes

# By default, the server modifies the process title (as seen in 'top' and 'ps') to

# provide some runtime information. It is possible to disable this and leave

# the process name as executed by setting the following to no.

set-proc-title yes

# When changing the process title, the server uses the following template to construct

# the modified title.

#

# Template variables are specified in curly brackets. The following variables are

# supported:

#

# {title}           Name of process as executed if parent, or type of child process.

# {listen-addr}     Bind address or '*' followed by TCP or TLS port listening on, or

#                   Unix socket if only that's available.

# {server-mode}     Special mode, i.e. "[sentinel]" or "[cluster]".

# {port}            TCP port listening on, or 0.

# {tls-port}        TLS port listening on, or 0.

# {unixsocket}      Unix domain socket listening on, or "".

# {config-file}     Name of configuration file used.

#

proc-title-template "{title} {listen-addr} {server-mode}"

# Set the local environment which is used for string comparison operations, and

# also affect the performance of Lua scripts. Empty String indicates the locale

# is derived from the environment variables.

locale-collate ""

# Valkey is largely compatible with Redis OSS, apart from a few cases where

# Valkey identifies itself as "Valkey" rather than "Redis". Extended

# Redis OSS compatibility mode makes Valkey pretend to be Redis. Enable this

# only if you have problems with tools or clients. This is a temporary

# configuration added in Valkey 8.0 and is scheduled to have no effect and

# will be removed in a future version.

#

# extended-redis-compatibility no

################################ SNAPSHOTTING  ################################

# Save the DB to disk.

#

# save <seconds> <changes> [<seconds> <changes> ...]

#

# The server will save the DB if the given number of seconds elapsed and it

# surpassed the given number of write operations against the DB.

#

# Snapshotting can be completely disabled with a single empty string argument

# as in following example:

#

# save ""

#

# Unless specified otherwise, by default the server will save the DB:

#   * After 3600 seconds (an hour) if at least 1 change was performed

#   * After 300 seconds (5 minutes) if at least 100 changes were performed

#   * After 60 seconds if at least 10000 changes were performed

#

# You can set these explicitly by uncommenting the following line.

#

# save 3600 1 300 100 60 10000

# By default the server will stop accepting writes if RDB snapshots are enabled

# (at least one save point) and the latest background save failed.

# This will make the user aware (in a hard way) that data is not persisting

# on disk properly, otherwise chances are that no one will notice and some

# disaster will happen.

#

# If the background saving process will start working again, the server will

# automatically allow writes again.

#

# However if you have setup your proper monitoring of the server

# and persistence, you may want to disable this feature so that the server will

# continue to work as usual even if there are problems with disk,

# permissions, and so forth.

stop-writes-on-bgsave-error yes

# Compress string objects using LZF when dump .rdb databases?

# By default compression is enabled as it's almost always a win.

# If you want to save some CPU in the saving child set it to 'no' but

# the dataset will likely be bigger if you have compressible values or keys.

rdbcompression yes

# Since version 5 of RDB a CRC64 checksum is placed at the end of the file.

# This makes the format more resistant to corruption but there is a performance

# hit to pay (around 10%) when saving and loading RDB files, so you can disable it

# for maximum performances.

#

# RDB files created with checksum disabled have a checksum of zero that will

# tell the loading code to skip the check.

rdbchecksum yes

# Valkey can try to load an RDB dump produced by a future version of Valkey.

# This can only work on a best-effort basis, because future RDB versions may

# contain information that's not known to the current version. If no new features

# are used, it may be possible to import the data produced by a later version,

# but loading is aborted if unknown information is encountered. Possible values

# are 'strict' and 'relaxed'. This also applies to replication and the RESTORE

# command.

rdb-version-check strict

# Enables or disables full sanitization checks for ziplist and listpack etc when

# loading an RDB or RESTORE payload. This reduces the chances of a assertion or

# crash later on while processing commands.

# Options:

#   no         - Never perform full sanitization

#   yes        - Always perform full sanitization

#   clients    - Perform full sanitization only for user connections.

#                Excludes: RDB files, RESTORE commands received from the primary

#                connection, and client connections which have the

#                skip-sanitize-payload ACL flag.

# The default should be 'clients' but since it currently affects cluster

# resharding via MIGRATE, it is temporarily set to 'no' by default.

#

# sanitize-dump-payload no

# The filename where to dump the DB

dbfilename dump.rdb

# Remove RDB files used by replication in instances without persistence

# enabled. By default this option is disabled, however there are environments

# where for regulations or other security concerns, RDB files persisted on

# disk by primaries in order to feed replicas, or stored on disk by replicas

# in order to load them for the initial synchronization, should be deleted

# ASAP. Note that this option ONLY WORKS in instances that have both AOF

# and RDB persistence disabled, otherwise is completely ignored.

#

# An alternative (and sometimes better) way to obtain the same effect is

# to use diskless replication on both primary and replicas instances. However

# in the case of replicas, diskless is not always an option.

rdb-del-sync-files no

# The working directory.

#

# The server log is written relative this directory, if the 'logfile'

# configuration directive is a relative path.

#

# The DB will be written inside this directory, with the filename specified

# above using the 'dbfilename' configuration directive.

#

# The Append Only File will also be created inside this directory.

#

# The Cluster config file is written relative this directory, if the

# 'cluster-config-file' configuration directive is a relative path.

#

# Note that you must specify a directory here, not a file name.

# Note that modifying 'dir' during runtime may have unexpected behavior,

# for example when a child process is running, related file operations may

# have unexpected effects.

dir ./

################################# REPLICATION #################################

# Master-Replica replication. Use replicaof to make a server a copy of

# another server. A few things to understand ASAP about replication.

#

#   +------------------+      +---------------+

#   |      Master      | ---> |    Replica    |

#   | (receive writes) |      |  (exact copy) |

#   +------------------+      +---------------+

#

# 1) Replication is asynchronous, but you can configure a primary to

#    stop accepting writes if it appears to be not connected with at least

#    a given number of replicas.

# 2) Replicas are able to perform a partial resynchronization with the

#    primary if the replication link is lost for a relatively small amount of

#    time. You may want to configure the replication backlog size (see the next

#    sections of this file) with a sensible value depending on your needs.

# 3) Replication is automatic and does not need user intervention. After a

#    network partition replicas automatically try to reconnect to primaries

#    and resynchronize with them.

#

# replicaof <primary_ip> <primary_port>

# If the primary is password protected (using the "requirepass" configuration

# directive below) it is possible to tell the replica to authenticate before

# starting the replication synchronization process, otherwise the primary will

# refuse the replica request.

#

# primaryauth <primary-password>

#

# However this is not enough if you are using ACLs

# and the default user is not capable of running the PSYNC

# command and/or other commands needed for replication. In this case it's

# better to configure a special user to use with replication, and specify the

# primaryuser configuration as such:

#

# primaryuser <username>

#

# When primaryuser is specified, the replica will authenticate against its

# primary using the new AUTH form: AUTH <username> <password>.

# When a replica loses its connection with the primary, or when the replication

# is still in progress, the replica can act in two different ways:

#

# 1) if replica-serve-stale-data is set to 'yes' (the default) the replica will

#    still reply to client requests, possibly with out of date data, or the

#    data set may just be empty if this is the first synchronization.

#

# 2) If replica-serve-stale-data is set to 'no' the replica will reply with error

#    "MASTERDOWN Link with MASTER is down and replica-serve-stale-data is set to 'no'"

#    to all data access commands, excluding commands such as:

#    INFO, REPLICAOF, AUTH, SHUTDOWN, REPLCONF, ROLE, CONFIG, SUBSCRIBE,

#    UNSUBSCRIBE, PSUBSCRIBE, PUNSUBSCRIBE, PUBLISH, PUBSUB, COMMAND, POST,

#    HOST and LATENCY.

#

replica-serve-stale-data yes

# You can configure a replica instance to accept writes or not. Writing against

# a replica instance may be useful to store some ephemeral data (because data

# written on a replica will be easily deleted after resync with the primary) but

# may also cause problems if clients are writing to it because of a

# misconfiguration.

#

# By default, replicas are read-only.

#

# Note: read only replicas are not designed to be exposed to untrusted clients

# on the internet. It's just a protection layer against misuse of the instance.

# Still a read only replica exports by default all the administrative commands

# such as CONFIG, DEBUG, and so forth. To a limited extent you can improve

# security of read only replicas using 'rename-command' to shadow all the

# administrative / dangerous commands.

replica-read-only yes

# Replication SYNC strategy: disk or socket.

#

# New replicas and reconnecting replicas that are not able to continue the

# replication process just receiving differences, need to do what is called a

# "full synchronization". An RDB file is transmitted from the primary to the

# replicas.

#

# The transmission can happen in two different ways:

#

# 1) Disk-backed: The primary creates a new process that writes the RDB

#                 file on disk. Later the file is transferred by the parent

#                 process to the replicas incrementally.

# 2) Diskless: The primary creates a new process that directly writes the

#              RDB file to replica sockets, without touching the disk at all.

#

# With disk-backed replication, while the RDB file is generated, more replicas

# can be queued and served with the RDB file as soon as the current child

# producing the RDB file finishes its work. With diskless replication instead

# once the transfer starts, new replicas arriving will be queued and a new

# transfer will start when the current one terminates.

#

# When diskless replication is used, the primary waits a configurable amount of

# time (in seconds) before starting the transfer in the hope that multiple

# replicas will arrive and the transfer can be parallelized.

#

# With slow disks and fast (large bandwidth) networks, diskless replication

# works better.

repl-diskless-sync yes

# When diskless replication is enabled, it is possible to configure the delay

# the server waits in order to spawn the child that transfers the RDB via socket

# to the replicas.

#

# This is important since once the transfer starts, it is not possible to serve

# new replicas arriving, that will be queued for the next RDB transfer, so the

# server waits a delay in order to let more replicas arrive.

#

# The delay is specified in seconds, and by default is 5 seconds. To disable

# it entirely just set it to 0 seconds and the transfer will start ASAP.

repl-diskless-sync-delay 5

# When diskless replication is enabled with a delay, it is possible to let

# the replication start before the maximum delay is reached if the maximum

# number of replicas expected have connected. Default of 0 means that the

# maximum is not defined and the server will wait the full delay.

repl-diskless-sync-max-replicas 0

# -----------------------------------------------------------------------------

# WARNING: Since in this setup the replica does not immediately store an RDB on

# disk, it may cause data loss during failovers. RDB diskless load + server

# modules not handling I/O reads may cause the server to abort in case of I/O errors

# during the initial synchronization stage with the primary.

# -----------------------------------------------------------------------------

#

# Replica can load the RDB it reads from the replication link directly from the

# socket, or store the RDB to a file and read that file after it was completely

# received from the primary.

#

# In many cases the disk is slower than the network, and storing and loading

# the RDB file may increase replication time (and even increase the primary's

# Copy on Write memory and replica buffers).

# However, when parsing the RDB file directly from the socket, in order to avoid

# data loss it's only safe to flush the current dataset when the new dataset is

# fully loaded in memory, resulting in higher memory usage.

# For this reason we have the following options:

#

# "disabled"          - Don't use diskless load (store the rdb file to the disk first)

# "swapdb"            - Keep current db contents in RAM while parsing the data directly

#                       from the socket. Replicas in this mode can keep serving current

#                       dataset while replication is in progress, except for cases where

#                       they can't recognize primary as having a data set from same

#                       replication history.

#                       Note that this requires sufficient memory, if you don't have it,

#                       you risk an OOM kill.

# "on-empty-db"       - Use diskless load only when current dataset is empty. This is

#                       safer and avoid having old and new dataset loaded side by side

#                       during replication.

# "flush-before-load" - [dangerous] Flush all data before parsing. Note that if

#                       there's a problem before the replication succeeded you may

#                       lose all your data.

repl-diskless-load disabled

# This dual channel replication sync feature optimizes the full synchronization process

# between a primary and its replicas. When enabled, it reduces both memory and CPU load

# on the primary server.

#

# How it works:

# 1. During full sync, instead of accumulating replication data on the primary server,

#    the data is sent directly to the syncing replica.

# 2. The primary's background save (bgsave) process streams the RDB snapshot directly

#    to the replica over a separate connection.

#

# Tradeoff:

# While this approach reduces load on the primary, it shifts the burden of storing

# the replication buffer to the replica. This means the replica must have sufficient

# memory to accommodate the buffer during synchronization. However, this tradeoff is

# generally beneficial as it prevents potential performance degradation on the primary

# server, which is typically handling more critical operations.

#

# When toggling this configuration on or off during an ongoing synchronization process,

# it does not change the already running sync method. The new configuration will take

# effect only for subsequent synchronization processes.

dual-channel-replication-enabled no

# Master send PINGs to its replicas in a predefined interval. It's possible to

# change this interval with the repl_ping_replica_period option. The default

# value is 10 seconds.

#

# repl-ping-replica-period 10

# The following option sets the replication timeout for:

#

# 1) Bulk transfer I/O during SYNC, from the point of view of replica.

# 2) Master timeout from the point of view of replicas (data, pings).

# 3) Replica timeout from the point of view of primaries (REPLCONF ACK pings).

#

# It is important to make sure that this value is greater than the value

# specified for repl-ping-replica-period otherwise a timeout will be detected

# every time there is low traffic between the primary and the replica. The default

# value is 60 seconds.

#

# repl-timeout 60

# Disable TCP_NODELAY on the replica socket after SYNC?

#

# If you select "yes", the server will use a smaller number of TCP packets and

# less bandwidth to send data to replicas. But this can add a delay for

# the data to appear on the replica side, up to 40 milliseconds with

# Linux kernels using a default configuration.

#

# If you select "no" the delay for data to appear on the replica side will

# be reduced but more bandwidth will be used for replication.

#

# By default we optimize for low latency, but in very high traffic conditions

# or when the primary and replicas are many hops away, turning this to "yes" may

# be a good idea.

repl-disable-tcp-nodelay no

# Enables MPTCP for the replica's connection to the primary.

#

# An MPTCP connection is established between the primary and the replica if

# the replica has set 'repl-mptcp yes' and the primary has set 'mptcp yes'.

# Otherwise, it will automatically and implicitly fall back to a regular TCP

# connection.

#

# repl-mptcp no

# Set the replication backlog size. The backlog is a buffer that accumulates

# replica data when replicas are disconnected for some time, so that when a

# replica wants to reconnect again, often a full resync is not needed, but a

# partial resync is enough, just passing the portion of data the replica

# missed while disconnected.

#

# The bigger the replication backlog, the longer the replica can endure the

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

#

# The backlog is only allocated if there is at least one replica connected.

#

# repl-backlog-size 10mb

# After a primary has no connected replicas for some time, the backlog will be

# freed. The following option configures the amount of seconds that need to

# elapse, starting from the time the last replica disconnected, for the backlog

# buffer to be freed.

#

# Note that replicas never free the backlog for timeout, since they may be

# promoted to primaries later, and should be able to correctly "partially

# resynchronize" with other replicas: hence they should always accumulate backlog.

#

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

#

# repl-backlog-ttl 3600

# The replica priority is an integer number published by the server in the INFO

# output. It is used by Sentinel in order to select a replica to promote

# into a primary if the primary is no longer working correctly.

#

# A replica with a low priority number is considered better for promotion, so

# for instance if there are three replicas with priority 10, 100, 25 Sentinel

# will pick the one with priority 10, that is the lowest.

#

# However a special priority of 0 marks the replica as not able to perform the

# role of primary, so a replica with priority of 0 will never be selected by

# Sentinel for promotion.

#

# By default the priority is 100.

replica-priority 100

# The propagation error behavior controls how the server will behave when it is

# unable to handle a command being processed in the replication stream from a primary

# or processed while reading from an AOF file. Errors that occur during propagation

# are unexpected, and can cause data inconsistency.

#

# If an application wants to ensure there is no data divergence, this configuration

# should be set to 'panic' instead. The value can also be set to 'panic-on-replicas'

# to only panic when a replica encounters an error on the replication stream. One of

# these two panic values will become the default value in the future once there are

# sufficient safety mechanisms in place to prevent false positive crashes.

#

# propagation-error-behavior ignore

# Replica ignore disk write errors controls the behavior of a replica when it is

# unable to persist a write command received from its primary to disk. By default,

# this configuration is set to 'no' and will crash the replica in this condition.

# It is not recommended to change this default.

#

# replica-ignore-disk-write-errors no

# Make the primary forbid expiration and eviction.

# This is useful for sync tools, because expiration and eviction may cause the data corruption.

# Sync tools can mark their connections as importing source by CLIENT IMPORT-SOURCE.

# NOTICE: Clients should avoid writing the same key on the source server and the destination server.

#

# import-mode no

# -----------------------------------------------------------------------------

# By default, Sentinel includes all replicas in its reports. A replica

# can be excluded from Sentinel's announcements. An unannounced replica

# will be ignored by the 'sentinel replicas <primary>' command and won't be

# exposed to Sentinel's clients.

#

# This option does not change the behavior of replica-priority. Even with

# replica-announced set to 'no', the replica can be promoted to primary. To

# prevent this behavior, set replica-priority to 0.

#

# replica-announced yes

# It is possible for a primary to stop accepting writes if there are less than

# N replicas connected, having a lag less or equal than M seconds.

#

# The N replicas need to be in "online" state.

#

# The lag in seconds, that must be <= the specified value, is calculated from

# the last ping received from the replica, that is usually sent every second.

#

# This option does not GUARANTEE that N replicas will accept the write, but

# will limit the window of exposure for lost writes in case not enough replicas

# are available, to the specified number of seconds.

#

# For example to require at least 3 replicas with a lag <= 10 seconds use:

#

# min-replicas-to-write 3

# min-replicas-max-lag 10

#

# Setting one or the other to 0 disables the feature.

#

# By default min-replicas-to-write is set to 0 (feature disabled) and

# min-replicas-max-lag is set to 10.

# A primary is able to list the address and port of the attached

# replicas in different ways. For example the "INFO replication" section

# offers this information, which is used, among other tools, by

# Sentinel in order to discover replica instances.

# Another place where this info is available is in the output of the

# "ROLE" command of a primary.

#

# The listed IP address and port normally reported by a replica is

# obtained in the following way:

#

#   IP: The address is auto detected by checking the peer address

#   of the socket used by the replica to connect with the primary.

#

#   Port: The port is communicated by the replica during the replication

#   handshake, and is normally the port that the replica is using to

#   listen for connections.

#

# However when port forwarding or Network Address Translation (NAT) is

# used, the replica may actually be reachable via different IP and port

# pairs. The following two options can be used by a replica in order to

# report to its primary a specific set of IP and port, so that both INFO

# and ROLE will report those values.

#

# There is no need to use both the options if you need to override just

# the port or the IP address.

#

# replica-announce-ip 5.5.5.5

# replica-announce-port 1234

############################### KEYS TRACKING #################################

# The client side caching of values is assisted via server-side support.

# This is implemented using an invalidation table that remembers, using

# a radix key indexed by key name, what clients have which keys. In turn

# this is used in order to send invalidation messages to clients. Please

# check this page to understand more about the feature:

#

#   https://valkey.io/topics/client-side-caching

#

# When tracking is enabled for a client, all the read only queries are assumed