一、背景与目的
测试与验证HA模式数据全量迁移到分片集群，
工具： redisShake4.3.1
二、准备
- 原集群
主： 172.25.242.158
从： 172.25.242.154
免密
SOX主机
端口：6379
redis-cli -h 172.25.242.158 -n 1 dbsize
2. 目的集群
#
172.25.39.54:6379
172.25.39.51:6379
172.25.39.52:6379
172.25.39.48:6379
172.25.39.55:6379
172.25.39.49:6379
172.25.39.53:6379
172.25.39.50:6379
组名： redis-cluster-newlps-ksa
密码： 2ghlmvl_ST
端口：6379
#
redis-cli -h 172.25.39.48 -a 2ghlmvl_MT cluster nodes
# 测试分片集群
172.25.190.13
172.25.189.209
172.25.189.211
172.25.190.12
172.25.190.11
172.25.189.212
密码
123456
redis-cli -h 172.25.190.13 -a 123456 dbsize
# redis-cli -h 172.25.189.209 -a 123456 dbsize
redis-cli -h 172.25.189.211 -a 123456 dbsize
redis-cli -h 172.25.190.12 -a 123456 dbsize
redis-cli -h 172.25.190.11 -a 123456 dbsize
redis-cli -h 172.25.189.212 -a 123456 dbsize
3. 工具
/root/redis/redisshake431
三、全量数据同步源
-
原集群数据量检查
[root@KSSYYVI01352 redisshake431]# redis-cli -h 172.25.242.158 -n 1 dbsize
(integer) 106992
[root@KSSYYVI01352 redisshake431]# redis-cli -h 172.25.242.158 -n 10 dbsize
(integer) 2486
2. ## redisShake配置文件调整shake.toml
[sync_reader]
cluster = false
address = "172.25.252.154:6379" # 源 Redis 地址
password = "" # 源 Redis 密码
sync_rdb = false # 启用 RDB 同步
sync_aof = false # 关闭 AOF 同步
[redis_writer]
cluster = true
address = "172.25.190.13:6379" # 目标 Redis 地址
password = "123456" # 目标 Redis 密码
skip_existing_keys = true
[filter]
allow_db = [01,10] # 只同步 1和10 号库的数据
#allow_keys = ["ycms:sm:draw:lottery:prize:current:count:2e2f3471006540194639:total"]
[advanced]
log_file = "shake.log" # 日志文件路径
log_level = "info" # 日志级别：info
pipeline_count_limit = 1024 # 默认管道大小
#rdb_restore_command_behavior = "skip" # 遇到重复键时覆盖
3. ## 执行
#
./redis-shake shake.toml
#
tail -f shake.log
shake.log 中显示如下：
执行的输出：
[root@KSSYYVI01352 redisshake431]# ./redis-shake shake.toml
2025-04-17 19:52:07 INF load config from file: shake.toml
2025-04-17 19:52:07 INF log_level: [info], log_file: [/root/redis/redisshake431/data/shake.log]
2025-04-17 19:52:07 INF changed work dir. dir=[/root/redis/redisshake431/data]
2025-04-17 19:52:07 INF GOMAXPROCS defaults to the value of runtime.NumCPU [4]
2025-04-17 19:52:07 INF not set pprof port
2025-04-17 19:52:07 INF create SyncStandaloneReader
2025-04-17 19:52:07 INF * address: 172.25.242.158:6379
2025-04-17 19:52:07 INF * username:
2025-04-17 19:52:07 INF * password:
2025-04-17 19:52:07 INF * tls: false
2025-04-17 19:52:07 INF create RedisClusterWriter
2025-04-17 19:52:07 INF * address (should be the address of one node in the Redis cluster): 172.25.190.13:6379
2025-04-17 19:52:07 INF * username:
2025-04-17 19:52:07 INF * password: ******
2025-04-17 19:52:07 INF * tls: false
2025-04-17 19:52:07 INF address=172.25.190.13:6379, reply=f99fe6a48492bd388474f33771a18b04fbf3c3d5 172.25.190.13:6379@16379 myself,master - 0 1744890726000 8 connected 5462-10922
c5815dabf9896b15ceda8dbd15742ab15d3e9fe6 172.25.190.11:6379@16379 slave 82a78b1cb3aaf11343704fb39f32e28f702fbfcc 0 1744890725039 6 connected
2150fe1f0122b2b03f2ab75ccd9516288800a2c8 172.25.189.211:6379@16379 master - 0 1744890727042 9 connected 10923-16383
82a78b1cb3aaf11343704fb39f32e28f702fbfcc 172.25.189.212:6379@16379 master - 0 1744890726039 6 connected 0-5461
5ec837e77676d921e135f7879c044786f7754b30 172.25.190.12:6379@16379 slave f99fe6a48492bd388474f33771a18b04fbf3c3d5 0 1744890725000 8 connected
e49b1fe2d24c3cd4452b32c9d78f99a2e9300c37 172.25.189.209:6379@16379 slave 2150fe1f0122b2b03f2ab75ccd9516288800a2c8 0 1744890725000 9 connected
2025-04-17 19:52:07 INF redisClusterWriter connected to redis cluster successful. addresses=[172.25.190.13:6379 172.25.189.211:6379 172.25.189.212:6379]
2025-04-17 19:52:07 INF start syncing...
2025-04-17 19:52:07 INF [reader_172.25.242.158_6379] source db is not doing bgsave! continue.
2025-04-17 19:52:11 ERR [writer_172.25.189.211_6379] receive reply failed. cmd=[select 1], error=[ERR SELECT is not allowed in cluster mode]
RedisShake/internal/writer/redis_standalone_writer.go:154 -> (*redisStandaloneWriter).processReply()
runtime/asm_amd64.s:1650 -> goexit()
说明:
主从模式向cluster迁移数据的时候，由于cluster模式不支持切换db，导致迁移失败。
cluster不支持select 选择db， 而源redis集群数据并不在db01中，所以失败。
建议：
如果源 Redis 使用了特定的数据库编号（如 1 或 10），而目标 Redis 集群不支持数据库编号，那么需要在源 Redis 中将数据迁移到默认数据库（通常是数据库 0）。
-
观察
#
redis-cli -h 172.25.190.13 -a 123456 dbsize
redis-cli -h 172.25.189.209 -a 123456 dbsize
redis-cli -h 172.25.189.211 -a 123456 dbsize
redis-cli -h 172.25.190.12 -a 123456 dbsize
redis-cli -h 172.25.190.11 -a 123456 dbsize
redis-cli -h 172.25.189.212 -a 123456 dbsize
-
对比新旧集群的数据量
# 新集群
redis-cli -h 172.25.39.54 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.51 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.52 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.48 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.55 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.49 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.53 -p 6379 -a 2ghlmvl_MT dbsize
redis-cli -h 172.25.39.50 -p 6379 -a 2ghlmvl_MT dbsize
-
完成
四、优化后的toml(toml中使用function实现db转换)
vi shake.toml
[sync_reader]
cluster = false # Set to true if the source is a Redis cluster
address = "172.25.242.158:6379" # For clusters, specify the address of any cluster node; use the master or slave address in master-slave mode
username = "" # Keep empty if ACL is not in use
password = "" # Keep empty if no authentication is required
tls = false # Set to true to enable TLS if needed
sync_rdb = true # Set to false if RDB synchronization is not required
sync_aof = false # Set to false if AOF synchronization is not required
prefer_replica = false # Set to true to sync from a replica node
try_diskless = false # Set to true for diskless sync if the source has repl-diskless-sync=yes
[redis_writer]
cluster = true # set to true if target is a redis cluster
address = "172.25.190.13:6379" # when cluster is true, set address to one of the cluster node
username = "" # keep empty if not using ACL
password = "123456" # keep empty if no authentication is required
tls = false
off_reply = false # turn off the server reply
[filter]
allow_keys = []
allow_key_prefix = []
allow_key_suffix = []
allow_key_regex = []
# Leave empty to block nothing
block_keys = []
block_key_prefix = []
block_key_suffix = []
block_key_regex = []
# Specify allowed and blocked database numbers (e.g., allow_db = [0, 1, 2], block_db = [3, 4, 5])
# Leave empty to allow all databases
allow_db = [1]
block_db = [0,2,3]
# Allow or block specific commands
# Leave empty to allow all commands
allow_command = []
block_command = []
# Allow or block specific command groups
# Leave empty to allow all command groups
allow_command_group = []
block_command_group = []
# Function for custom data processing
function = """
shake.log(DB)
local db1 = 1
local db2 = 0
if DB == db1 then
DB = db2
end
shake.call(DB, ARGV)
"""
[advanced]
dir = "data"
ncpu = 0 # runtime.GOMAXPROCS, 0 means use runtime.NumCPU() cpu cores
pprof_port = 0 # pprof port, 0 means disable
status_port = 0 # status port, 0 means disable
# log
log_file = "shake.log"
log_level = "info" # debug, info or warn
log_interval = 5 # in seconds
log_rotation = true # enable log rotation
log_max_size = 512 # MiB, logs max size to rotate, default 512 MiB
log_max_age = 7 # days, logs are kept, default 7 days
log_max_backups = 3 # number of log backups, default 3
log_compress = true # enable log compression after rotate, default true
# redis-shake gets key and value from rdb file, and uses RESTORE command to
# create the key in target redis. Redis RESTORE will return a "Target key name
# is busy" error when key already exists. You can use this configuration item
# to change the default behavior of restore:
# panic: redis-shake will stop when meet "Target key name is busy" error.
# rewrite: redis-shake will replace the key with new value.
# skip: redis-shake will skip restore the key when meet "Target key name is busy" error.
rdb_restore_command_behavior = "panic" # panic, rewrite or skip
# redis-shake uses pipeline to improve sending performance.
# Adjust this value based on the destination Redis performance:
# - Higher values may improve performance for capable destinations.
# - Lower values are recommended for destinations with poor performance.
# 1024 is a good default value for most cases.
pipeline_count_limit = 1024
# This setting corresponds to the 'client-query-buffer-limit' in Redis configuration.
# The default value is typically 1GB.
# It's recommended not to modify this value unless absolutely necessary.
target_redis_client_max_querybuf_len = 1073741824 # 1GB in bytes
# This setting corresponds to the 'proto-max-bulk-len' in Redis configuration.
# It defines the maximum size of a single string element in the Redis protocol.
# The value must be 1MB or greater. Default is 512MB.
# It's recommended not to modify this value unless absolutely necessary.
target_redis_proto_max_bulk_len = 512_000_000
# If the source is Elasticache, you can set this item. AWS ElastiCache has custom
# psync command, which can be obtained through a ticket.
aws_psync = "" # example: aws_psync = "10.0.0.1:6379@nmfu2sl5osync,10.0.0.1:6379@xhma21xfkssync"
# destination will delete itself entire database before fetching files
# from source during full synchronization.
# This option is similar redis replicas RDB diskless load option:
# repl-diskless-load on-empty-db
empty_db_before_sync = false
[module]
# The data format for BF.LOADCHUNK is not compatible in different versions. v2.6.3 <=> 20603
target_mbbloom_version = 20603
调用方式
./redis-shake shake.toml
说明： function函数实现转换规则，将源redis中的db1中数据迁移到cluster中
# Function for custom data processing
function = """
shake.log(DB)
local db1 = 1
local db2 = 0
if DB == db1 then
DB = db2
end
shake.call(DB, ARGV)
"""
执行后的输出日志
五、补充信息
- 文档与参考
tair-opensource.github.io/RedisShake/…
-
相关方案 lue
-
注意事项 · 不要在同一个目录运行两个RedisShake进程,因为运行时产生的临时文件可能会被覆盖,导 致异常行为。 · 不要降低Redis版本,比如从6.0降到5.0,因为RedisShake每个大版本都会引入一些新的 命令和新的编码方式,如果降低版本,可能会导致不兼容。
