path: root/tikv.toml

# TiKV config template
#  Human-readable big numbers:
#   File size(based on byte): KB, MB, GB, TB, PB
#    e.g.: 1_048_576 = "1MB"
#   Time(based on ms): ms, s, m, h
#    e.g.: 78_000 = "1.3m"

# log level: trace, debug, info, warning, error, critical.
# Note that `debug` and `trace` are only available in development builds.
# log-level = "info"

# file to store log, write to stderr if it's empty.
# log-file = ""

# timespan between rotating the log files.
# Once this timespan passes the existing log file will have a timestamp appended to its name,
# and a new file will be created.
# log-rotation-timespan = "24h"

[readpool.storage]
# size of thread pool for high-priority operations
# high-concurrency = 4
# size of thread pool for normal-priority operations
# normal-concurrency = 4
# size of thread pool for low-priority operations
# low-concurrency = 4
# max running high-priority operations of each worker, reject if exceed
# max-tasks-per-worker-high = 2000
# max running normal-priority operations of each worker, reject if exceed
# max-tasks-per-worker-normal = 2000
# max running low-priority operations of each worker, reject if exceed
# max-tasks-per-worker-low = 2000
# size of stack size for each thread pool
# stack-size = "10MB"

[readpool.coprocessor]
# Notice: if CPU_NUM > 8, default thread pool size for coprocessors
# will be set to CPU_NUM * 0.8.

# high-concurrency = 8
# normal-concurrency = 8
# low-concurrency = 8
# max-tasks-per-worker-high = 2000
# max-tasks-per-worker-normal = 2000
# max-tasks-per-worker-low = 2000
# stack-size = "10MB"

[server]
# set listening address.
# addr = "127.0.0.1:20160"
# set advertise listening address for client communication, if not set, use addr instead.
# advertise-addr = ""
# notify capacity, 40960 is suitable for about 7000 regions.
# notify-capacity = 40960
# maximum number of messages can be processed in one tick.
# messages-per-tick = 4096

## Status address.
## This is used for reporting the status of TiKV directly through the HTTP address.
## Empty string means disabling it.
# status-addr = "127.0.0.1:20180"

## Set the maximum number of worker threads for the status report HTTP service.
# status-thread-pool-size = 1

## Compression type for gRPC channel: none, deflate or gzip.
# grpc-compression-type = "none"
# size of thread pool for grpc server.
# grpc-concurrency = 4
# The number of max concurrent streams/requests on a client connection.
# grpc-concurrent-stream = 1024
# The number of connections with each tikv server to send raft messages.
# grpc-raft-conn-num = 10
# Amount to read ahead on individual grpc streams.
# grpc-stream-initial-window-size = "2MB"
# Time to wait before sending out a ping to check if server is still alive.
# This is only for communications between tikv instances.
# grpc-keepalive-time = "10s"
# Time to wait before closing the connection without receiving keepalive ping
# ack.
# grpc-keepalive-timeout = "3s"

# How many snapshots can be sent concurrently.
# concurrent-send-snap-limit = 32
# How many snapshots can be recv concurrently.
# concurrent-recv-snap-limit = 32

# max recursion level allowed when decoding dag expression
# end-point-recursion-limit = 1000

# max time to handle coprocessor request before timeout
# end-point-request-max-handle-duration = "60s"

# the max bytes that snapshot can be written to disk in one second,
# should be set based on your disk performance
# snap-max-write-bytes-per-sec = "100MB"

# set attributes about this server, e.g. { zone = "us-west-1", disk = "ssd" }.
# labels = {}

[storage]
# set the path to rocksdb directory.
data-dir = "/var/lib/tikv/store"

# notify capacity of scheduler's channel
# scheduler-notify-capacity = 10240

# the number of slots in scheduler latches, concurrency control for write.
# scheduler-concurrency = 2048000

# scheduler's worker pool size, should increase it in heavy write cases,
# also should less than total cpu cores.
# scheduler-worker-pool-size = 4

# When the pending write bytes exceeds this threshold,
# the "scheduler too busy" error is displayed.
# scheduler-pending-write-threshold = "100MB"

[pd]
# pd endpoints
endpoints = ["127.0.0.1:2379"]

[raftstore]
# true (default value) for high reliability, this can prevent data loss when power failure.
# sync-log = true

# minimizes disruption when a partitioned node rejoins the cluster by using a two phase election.
# prevote = true

# set the path to raftdb directory, default value is data-dir/raft
# raftdb-path = ""

# set store capacity, if no set, use disk capacity.
# capacity = 0

# notify capacity, 40960 is suitable for about 7000 regions.
# notify-capacity = 40960

# maximum number of messages can be processed in one tick.
# messages-per-tick = 4096

# Region heartbeat tick interval for reporting to pd.
# pd-heartbeat-tick-interval = "60s"
# Store heartbeat tick interval for reporting to pd.
# pd-store-heartbeat-tick-interval = "10s"

# When region size changes exceeds region-split-check-diff, we should check
# whether the region should be split or not.
# region-split-check-diff = "6MB"

# Interval to check region whether need to be split or not.
# split-region-check-tick-interval = "10s"

# When raft entry exceed the max size, reject to propose the entry.
# raft-entry-max-size = "8MB"

# Interval to gc unnecessary raft log.
# raft-log-gc-tick-interval = "10s"
# A threshold to gc stale raft log, must >= 1.
# raft-log-gc-threshold = 50
# When entry count exceed this value, gc will be forced trigger.
# raft-log-gc-count-limit = 72000
# When the approximate size of raft log entries exceed this value, gc will be forced trigger.
# It's recommanded to set it to 3/4 of region-split-size.
# raft-log-gc-size-limit = "72MB"

# When a peer hasn't been active for max-peer-down-duration,
# we will consider this peer to be down and report it to pd.
# max-peer-down-duration = "5m"

# Interval to check whether start manual compaction for a region,
# region-compact-check-interval = "5m"

# Number of regions for each time to check.
# region-compact-check-step = 100

# The minimum number of delete tombstones to trigger manual compaction.
# region-compact-min-tombstones = 10000

# The minimum percentage of delete tombstones to trigger manual compaction.
# Should between 1 and 100. Manual compaction only triggered when the number
# of delete tombstones exceeds region-compact-min-tombstones and the percentage
# of delete tombstones exceeds region-compact-tombstones-percent.
# region-compact-tombstones-percent = 30

# Interval to check whether should start a manual compaction for lock column family,
# if written bytes reach lock-cf-compact-threshold for lock column family, will fire
# a manual compaction for lock column family.
# lock-cf-compact-interval = "10m"

# lock-cf-compact-bytes-threshold = "256MB"

# Interval (s) to check region whether the data are consistent.
# consistency-check-interval = 0

# Use delete range to drop a large number of continuous keys.
# use-delete-range = false

# delay time before deleting a stale peer
# clean-stale-peer-delay = "10m"

# Interval to cleanup import sst files.
# cleanup-import-sst-interval = "10m"

[coprocessor]
# When it is true, it will try to split a region with table prefix if
# that region crosses tables. It is recommended to turn off this option
# if there will be a large number of tables created.
# split-region-on-table = true
# When the region's size exceeds region-max-size, we will split the region
# into two which the left region's size will be region-split-size or a little
# bit smaller.
# region-max-size = "144MB"
# region-split-size = "96MB"
# When the region's keys exceeds region-max-keys, we will split the region
# into two which the left region's keys will be region-split-keys or a little
# bit smaller.
# region-max-keys = 1440000
# region-split-keys = 960000

[rocksdb]
# Maximum number of concurrent background jobs (compactions and flushes)
# max-background-jobs = 8

# This value represents the maximum number of threads that will concurrently perform a
# compaction job by breaking it into multiple, smaller ones that are run simultaneously.
# Default: 1 (i.e. no subcompactions)
# max-sub-compactions = 1

# Number of open files that can be used by the DB.  You may need to
# increase this if your database has a large working set. Value -1 means
# files opened are always kept open. You can estimate number of files based
# on target_file_size_base and target_file_size_multiplier for level-based
# compaction.
# If max-open-files = -1, RocksDB will prefetch index and filter blocks into
# block cache at startup, so if your database has a large working set, it will
# take several minutes to open the db.
# max-open-files = 40960

# Max size of rocksdb's MANIFEST file.
# For detailed explanation please refer to https://github.com/facebook/rocksdb/wiki/MANIFEST
# max-manifest-file-size = "128MB"

# If true, the database will be created if it is missing.
# create-if-missing = true

# rocksdb wal recovery mode
# 0 : TolerateCorruptedTailRecords, tolerate incomplete record in trailing data on all logs;
# 1 : AbsoluteConsistency, We don't expect to find any corruption in the WAL;
# 2 : PointInTimeRecovery, Recover to point-in-time consistency;
# 3 : SkipAnyCorruptedRecords, Recovery after a disaster;
# wal-recovery-mode = 2

# rocksdb write-ahead logs dir path
# This specifies the absolute dir path for write-ahead logs (WAL).
# If it is empty, the log files will be in the same dir as data.
# When you set the path to rocksdb directory in memory like in /dev/shm, you may want to set
# wal-dir to a directory on a persistent storage.
# See https://github.com/facebook/rocksdb/wiki/How-to-persist-in-memory-RocksDB-database
# wal-dir = "/tmp/tikv/store"

# The following two fields affect how archived write-ahead logs will be deleted.
# 1. If both set to 0, logs will be deleted asap and will not get into the archive.
# 2. If wal-ttl-seconds is 0 and wal-size-limit is not 0,
#    WAL files will be checked every 10 min and if total size is greater
#    then wal-size-limit, they will be deleted starting with the
#    earliest until size_limit is met. All empty files will be deleted.
# 3. If wal-ttl-seconds is not 0 and wal-size-limit is 0, then
#    WAL files will be checked every wal-ttl-seconds / 2 and those that
#    are older than wal-ttl-seconds will be deleted.
# 4. If both are not 0, WAL files will be checked every 10 min and both
#    checks will be performed with ttl being first.
# When you set the path to rocksdb directory in memory like in /dev/shm, you may want to set
# wal-ttl-seconds to a value greater than 0 (like 86400) and backup your db on a regular basis.
# See https://github.com/facebook/rocksdb/wiki/How-to-persist-in-memory-RocksDB-database
# wal-ttl-seconds = 0
# wal-size-limit = 0

# rocksdb max total wal size
# max-total-wal-size = "4GB"

# Rocksdb Statistics provides cumulative stats over time.
# Turn statistics on will introduce about 5%-10% overhead for RocksDB,
# but it is worthy to know the internal status of RocksDB.
# enable-statistics = true

# Dump statistics periodically in information logs.
# Same as rocksdb's default value (10 min).
# stats-dump-period = "10m"

# Due to Rocksdb FAQ: https://github.com/facebook/rocksdb/wiki/RocksDB-FAQ,
# If you want to use rocksdb on multi disks or spinning disks, you should set value at
# least 2MB;
# compaction-readahead-size = 0

# This is the maximum buffer size that is used by WritableFileWrite
# writable-file-max-buffer-size = "1MB"

# Use O_DIRECT for both reads and writes in background flush and compactions
# use-direct-io-for-flush-and-compaction = false

# Limit the disk IO of compaction and flush. Compaction and flush can cause
# terrible spikes if they exceed a certain threshold. Consider setting this to
# 50% ~ 80% of the disk throughput for a more stable result. However, in heavy
# write workload, limiting compaction and flush speed can cause write stalls too.
# rate-bytes-per-sec = 0

# Enable or disable the pipelined write
# enable-pipelined-write = true

# Allows OS to incrementally sync files to disk while they are being
# written, asynchronously, in the background.
# bytes-per-sync = "1MB"

# Allows OS to incrementally sync WAL to disk while it is being written.
# wal-bytes-per-sync = "512KB"

# Specify the maximal size of the Rocksdb info log file. If the log file
# is larger than `max_log_file_size`, a new info log file will be created.
# If max_log_file_size == 0, all logs will be written to one log file.
# info-log-max-size = "1GB"

# Time for the Rocksdb info log file to roll (in seconds).
# If specified with non-zero value, log file will be rolled
# if it has been active longer than `log_file_time_to_roll`.
# 0 means disabled.
# info-log-roll-time = "0"

# Maximal Rocksdb info log files to be kept.
# info-log-keep-log-file-num = 10

# This specifies the Rocksdb info LOG dir.
# If it is empty, the log files will be in the same dir as data.
# If it is non empty, the log files will be in the specified dir,
# and the db data dir's absolute path will be used as the log file
# name's prefix.
# info-log-dir = ""

# Column Family default used to store actual data of the database.
[rocksdb.defaultcf]
# compression method (if any) is used to compress a block.
#   no:     kNoCompression
#   snappy: kSnappyCompression
#   zlib:   kZlibCompression
#   bzip2:  kBZip2Compression
#   lz4:    kLZ4Compression
#   lz4hc:  kLZ4HCCompression
#   zstd:   kZSTD

# per level compression
# compression-per-level = ["no", "no", "lz4", "lz4", "lz4", "zstd", "zstd"]

# Approximate size of user data packed per block.  Note that the
# block size specified here corresponds to uncompressed data.
# block-size = "64KB"

# If you're doing point lookups you definitely want to turn bloom filters on, We use
# bloom filters to avoid unnecessary disk reads. Default bits_per_key is 10, which
# yields ~1% false positive rate. Larger bits_per_key values will reduce false positive
# rate, but increase memory usage and space amplification.
# bloom-filter-bits-per-key = 10

# false means one sst file one bloom filter, true means evry block has a corresponding bloom filter
# block-based-bloom-filter = false

# level0-file-num-compaction-trigger = 4

# Soft limit on number of level-0 files. We start slowing down writes at this point.
# level0-slowdown-writes-trigger = 20

# Maximum number of level-0 files.  We stop writes at this point.
# level0-stop-writes-trigger = 36

# Amount of data to build up in memory (backed by an unsorted log
# on disk) before converting to a sorted on-disk file.
# write-buffer-size = "128MB"

# The maximum number of write buffers that are built up in memory.
# max-write-buffer-number = 5

# The minimum number of write buffers that will be merged together
# before writing to storage.
# min-write-buffer-number-to-merge = 1

# Control maximum total data size for base level (level 1).
# max-bytes-for-level-base = "512MB"

# Target file size for compaction.
# target-file-size-base = "8MB"

# Max bytes for compaction.max_compaction_bytes
# max-compaction-bytes = "2GB"

# There are four different algorithms to pick files to compact.
# 0 : ByCompensatedSize
# 1 : OldestLargestSeqFirst
# 2 : OldestSmallestSeqFirst
# 3 : MinOverlappingRatio
# compaction-pri = 3

# block-cache used to cache uncompressed blocks, big block-cache can speed up read.
# in normal cases should tune to 30%-50% system's total memory.
# block-cache-size = "1GB"

# Indicating if we'd put index/filter blocks to the block cache.
# If not specified, each "table reader" object will pre-load index/filter block
# during table initialization.
# cache-index-and-filter-blocks = true

# Pin level0 filter and index blocks in cache.
# pin-l0-filter-and-index-blocks = true

# Enable read amplication statistics.
# value  =>  memory usage (percentage of loaded blocks memory)
# 1      =>  12.50 %
# 2      =>  06.25 %
# 4      =>  03.12 %
# 8      =>  01.56 %
# 16     =>  00.78 %
# read-amp-bytes-per-bit = 0

# Pick target size of each level dynamically.
# dynamic-level-bytes = true

# Options for Column Family write
# Column Family write used to store commit informations in MVCC model
[rocksdb.writecf]
# compression-per-level = ["no", "no", "lz4", "lz4", "lz4", "zstd", "zstd"]
# block-size = "64KB"
# write-buffer-size = "128MB"
# max-write-buffer-number = 5
# min-write-buffer-number-to-merge = 1
# max-bytes-for-level-base = "512MB"
# target-file-size-base = "8MB"

# in normal cases should tune to 10%-30% system's total memory.
# block-cache-size = "256MB"
# level0-file-num-compaction-trigger = 4
# level0-slowdown-writes-trigger = 20
# level0-stop-writes-trigger = 36
# cache-index-and-filter-blocks = true
# pin-l0-filter-and-index-blocks = true
# compaction-pri = 3
# read-amp-bytes-per-bit = 0
# dynamic-level-bytes = true

[rocksdb.lockcf]
# compression-per-level = ["no", "no", "no", "no", "no", "no", "no"]
# block-size = "16KB"
# write-buffer-size = "128MB"
# max-write-buffer-number = 5
# min-write-buffer-number-to-merge = 1
# max-bytes-for-level-base = "128MB"
# target-file-size-base = "8MB"
# block-cache-size = "256MB"
# level0-file-num-compaction-trigger = 1
# level0-slowdown-writes-trigger = 20
# level0-stop-writes-trigger = 36
# cache-index-and-filter-blocks = true
# pin-l0-filter-and-index-blocks = true
# compaction-pri = 0
# read-amp-bytes-per-bit = 0
# dynamic-level-bytes = true

[raftdb]
# max-sub-compactions = 1
# max-open-files = 40960
# max-manifest-file-size = "20MB"
# create-if-missing = true

# enable-statistics = true
# stats-dump-period = "10m"

# compaction-readahead-size = 0
# writable-file-max-buffer-size = "1MB"
# use-direct-io-for-flush-and-compaction = false
# enable-pipelined-write = true
# allow-concurrent-memtable-write = false
# bytes-per-sync = "1MB"
# wal-bytes-per-sync = "512KB"

# info-log-max-size = "1GB"
# info-log-roll-time = "0"
# info-log-keep-log-file-num = 10
# info-log-dir = ""

[raftdb.defaultcf]
# compression-per-level = ["no", "no", "lz4", "lz4", "lz4", "zstd", "zstd"]
# block-size = "64KB"
# write-buffer-size = "128MB"
# max-write-buffer-number = 5
# min-write-buffer-number-to-merge = 1
# max-bytes-for-level-base = "512MB"
# target-file-size-base = "8MB"

# should tune to 256MB~2GB.
# block-cache-size = "256MB"
# level0-file-num-compaction-trigger = 4
# level0-slowdown-writes-trigger = 20
# level0-stop-writes-trigger = 36
# cache-index-and-filter-blocks = true
# pin-l0-filter-and-index-blocks = true
# compaction-pri = 0
# read-amp-bytes-per-bit = 0
# dynamic-level-bytes = true

[security]
# set the path for certificates. Empty string means disabling secure connectoins.
# ca-path = ""
# cert-path = ""
# key-path = ""

[import]
# number of threads to handle RPC requests.
# num-threads = 8
# stream channel window size, stream will be blocked on channel full.
# stream-channel-window = 128