-
trocks secondary development command
Command Desc hsetex hset with expire hmsetex hmset with expire -
Sortedint Commands（ROR recommends using redis’ set commands, such as scard, sadd...）
Command Desc sicard like scard siadd like sadd, but member is int sirem like srem, but member is int sirange sirange key offset count cursor since_id sirevrange sirevrange key offset count cursor max_id -
LOCK Commands (ROR recommends using multi transactions instead)
Command Desc getl lock key, act is similar with "get" setfl release key, act is similar with "set" abortfl release key, quit transaction anyway setflex release key. act is similar with "setex" showl monitor lock actions freel release key, quit transaction by condition hgetl lock key, act is similar with "hget" hsetfl release key, act is similar with "hset" hsetflex release key, act is similar with "hsetex" hmgetl lock key, act is similar with "hmget" hmsetfl release key, act is similar with "hmset" hmsetflex release key, act is similar with "hmsetex" -
Commands supported by ROR
-
Redis native commands
Command Type Whether to support Supported version Remark get string √ 1.0.0 getex string √ 1.0.0 getdel string √ 1.0.0 set string √ 1.0.0 setnx string √ 1.0.0 setex string √ 1.0.0 psetex string √ 1.0.0 append string √ 1.0.0 strlen string √ 1.0.0 del string √ 1.0.0 unlink string √ 1.0.0 exists string √ 1.0.0 setbit string √ 1.0.0 getbit string √ 1.0.0 bitfield string √ 1.0.0 bitfield_ro string √ 1.0.0 setrange string √ 1.0.0 getrange string √ 1.0.0 substr string √ 1.0.0 incr string √ 1.0.0 decr string √ 1.0.0 mget string √ 1.0.0 incrby string √ 1.0.0 decrby string √ 1.0.0 incrbyfloat string √ 1.0.0 getset string √ 1.0.0 mset string √ 1.0.0 msetnx string √ 1.0.0 pfselftest string √ 1.0.0 pfadd string √ 1.0.0 pfcount string √ 1.0.0 pfmerge string √ 1.0.0 pfdebug string √ 1.0.0 stralgo string √ 1.0.0 bitop string √ 1.0.0 bitcount string √ 1.0.0 bitpos string √ 1.0.0 rpush list √ 1.0.0 lpush list √ 1.0.0 rpushx list √ 1.0.0 lpushx list √ 1.0.0 linsert list √ 1.0.0 rpop list √ 1.0.0 lpop list √ 1.0.0 brpop list √ 1.0.0 brpoplpush list √ 1.0.3 blmove list √ 1.0.3 blpop list √ 1.0.0 llen list √ 1.0.0 lindex list √ 1.0.0 lset list √ 1.0.0 lrange list √ 1.0.0 ltrim list √ 1.0.0 lpos list √ 1.0.0 lrem list √ 1.0.0 rpoplpush list √ 1.0.0 lmove list √ 1.0.0 sadd set √ 1.0.0 srem set √ 1.0.0 smove set √ 1.0.0 sismember set √ 1.0.0 smismember set √ 1.0.0 scard set √ 1.0.0 spop set √ 1.0.0 srandmember set √ 1.0.0 sinter set √ 1.0.0 sinterstore set √ 1.0.0 sunion set √ 1.0.0 sunionstore set √ 1.0.0 sdiff set √ 1.0.0 sdiffstore set √ 1.0.0 smembers set √ 1.0.0 sscan set √ 1.0.0 zadd zset √ 1.0.0 zincrby zset √ 1.0.0 zrem zset √ 1.0.0 zremrangebyscore zset √ 1.0.0 zremrangebyrank zset √ 1.0.0 zremrangebylex zset √ 1.0.0 zunionstore zset √ 1.0.0 zinterstore zset √ 1.0.0 zdiffstore zset √ 1.0.0 zunion zset √ 1.0.0 zinter zset √ 1.0.0 zdiff zset √ 1.0.0 zrange zset √ 1.0.0 zrangestore zset √ 1.0.0 zrangebyscore zset √ 1.0.0 zrevrangebyscore zset √ 1.0.0 zrangebylex zset √ 1.0.0 zrevrangebylex zset √ 1.0.0 zcount zset √ 1.0.0 zlexcount zset √ 1.0.0 zrevrange zset √ 1.0.0 zcard zset √ 1.0.0 zscore zset √ 1.0.0 zmscore zset √ 1.0.0 zrank zset √ 1.0.0 zrevrank zset √ 1.0.0 zscan zset √ 1.0.0 zpopmin zset √ 1.0.0 zpopmax zset √ 1.0.0 bzpopmin zset √ 1.0.0 bzpopmax zset √ 1.0.0 zrandmember zset √ 1.0.0 hset hash √ 1.0.0 hsetnx hash √ 1.0.0 hget hash √ 1.0.0 hmset hash √ 1.0.0 hmget hash √ 1.0.0 hincrby hash √ 1.0.0 hincrbyfloat hash √ 1.0.0 hdel hash √ 1.0.0 hlen hash √ 1.0.0 hstrlen hash √ 1.0.0 hkeys hash √ 1.0.0 hvals hash √ 1.0.0 hgetall hash √ 1.0.0 hexists hash √ 1.0.0 hrandfield hash √ 1.0.0 hscan hash √ 1.0.0 dbsize keyspace √ 1.0.0 randomkey keyspace √ 1.0.0 expire keyspace √ 1.0.0 expireat keyspace √ 1.0.0 pexpire keyspace √ 1.0.0 pexpireat keyspace √ 1.0.0 ttl keyspace √ 1.0.0 touch keyspace √ 1.0.0 pttl keyspace √ 1.0.0 persist keyspace √ 1.0.0 type keyspace √ 1.0.0 scan keyspace √* 1.0.0 The cursor must start from 0, and the next cursor must be the cursor returned by ROR last time (arbitrary starting of the cursor is not supported). In other words, if the client disconnects and reconnects, the cursor must start from 0 again. dump keyspace √ 1.0.0 object keyspace √ 1.0.0 flushdb keyspace √ 1.0.0 flushall keyspace √ 1.0.0 select keyspace √* 1.0.0 Select in eval is not supported (but select in multi/exec is supported) move keyspace √ 1.0.0 copy keyspace √ 1.0.0 rename keyspace √ 1.0.0 renamenx keyspace √ 1.0.0 keys keyspace × swapdb keyspace × auth server √ 1.0.0 ping server √ 1.0.0 echo server √ 1.0.0 save server √ 1.0.0 bgsave server √ 1.0.0 refullsync server √ 1.0.0 bgrewriteaof server √ 1.0.0 shutdown server √ 1.0.0 lastsave server √ 1.0.0 multi transaction √ 1.0.0 exec transaction √ 1.0.0 discard transaction √ 1.0.0 watch transaction √* 1.0.0 The impact of the flushdb command on watch Redis and ROR behave differently:
* Redis: If the key to be watched does not exist in the db, flushdb will not cause subsequent exec to terminate.
* ROR: Even if the watched key does not exist in the db, flushdb will cause subsequent exec to fail.
The main reason is: ror does not contain the full amount of keys, flushdb cannot determine whether the watched key does not exist, and handles it according to the existence of the key, terminating the current transaction with the watch key.unwatch transaction √ 1.0.0 eval transaction √ 1.0.0 evalsha transaction √ 1.0.0 subscribe pubsub √ 1.0.0 unsubscribe pubsub √ 1.0.0 psubscribe pubsub √ 1.0.0 punsubscribe pubsub √ 1.0.0 publish pubsub √ 1.0.0 pubsub pubsub √ 1.0.0 cluster cluster √ 1.0.0 restore cluster √ 1.0.0 restore-asking cluster √ 1.0.0 migrate cluster √ 1.0.0 asking cluster √ 1.0.0 readonly cluster √ 1.0.0 readwrite cluster √ 1.0.0 geoadd geo √ 1.0.0 georadius geo √ 1.0.0 georadius_ro geo √ 1.0.0 georadiusbymember geo √ 1.0.0 georadiusbymember_ro geo √ 1.0.0 geohash geo √ 1.0.0 geopos geo √ 1.0.0 geodist geo √ 1.0.0 geosearch geo √ 1.0.0 geosearchstore geo √ 1.0.0 xadd stream √ 1.0.0 xrange stream √ 1.0.0 xrevrange stream √ 1.0.0 xlen stream √ 1.0.0 xread stream √ 1.0.0 xreadgroup stream √ 1.0.0 xgroup stream √ 1.0.0 xsetid stream √ 1.0.0 xack stream √ 1.0.0 xpending stream √ 1.0.0 xclaim stream √ 1.0.0 xautoclaim stream √ 1.0.0 xinfo stream √ 1.0.0 xdel stream √ 1.0.0 xtrim stream √ 1.0.0 wait server √ 1.0.0 command server √ 1.0.0 latency server √ 1.0.0 acl server √ 1.0.0 reset server √ 1.0.0 failover server √ 1.0.0 sync server √ 1.0.0 psync server √ 1.0.0 replconf server √ 1.0.0 sort server √ 1.0.0 info server √ 1.0.0 monitor server √ 1.0.0 slaveof server √ 1.0.0 xslaveof server √ 1.0.0 replicaof server √ 1.0.0 role server √ 1.0.0 debug server √ 1.0.0 config server √ 1.0.0 client server √ 1.0.0 hello server √ 1.0.0 slowlog server √ 1.0.0 script server √ 1.0.0 time server √ 1.0.0 memory server √ 1.0.0 -
SWAP Related commands
Command Supported version Grammer Description Remark swap 1.0.1 1) SWAP [<arg> [value] [opt] ...]. Subcommands are:
2) OBJECT
3) Show info aboutkeyand assosiated value.
4) ENCODE-META-KEY
5) Encode meta key.
6) DECODE-META-KEY
7) Decode meta key.
8) ENCODE-DATA-KEY
9) Encode data key.
10) DECODE-DATA-KEY
11) Decode data key.
12) RIO-GET meta|data <rawkey> <rawkey> ...
13) Get raw value from rocksdb.
14) RIO-SCAN meta|data <prefix>
15) Scan rocksdb with prefix.
16) RIO-ERROR <count>
17) Make next count rio return error.
18) RESET-STATS
19) Reset swap stats.
20) COMPACT
21) COMPACT rocksdb
22) ROCKSDB-PROPERTY-INT [<cfname,cfname...>]
23) Get rocksdb property value (int type)
24) ROCKSDB-PROPERTY-VALUE [<cfname,cfname...>]
25) Get rocksdb property value (string type)
26) HELP
27) Prints this help.Display information about keys and associated values swap function auxiliary command swap.evict 1.0.1 swap.evict key1 key2 ... Internal command that can be used to manually remove keys swap.slowlog 1.0.1 1) SWAP.SLOWLOG [<arg> [value] [opt] ...]. Subcommands are:
2) GET [<count>]
3) Return top <count> entries from the slowlog (default: 10). Entries are
4) made of:
5) id, timestamp, time in microseconds, swap cnt, swap time in microseconds, arguments array,
6) client IP and port, client nasme,
7) swap debug traces(need open swap-debug-trace-latency before).
8) HELP
9) Prints this help.slowlog displays additional swap-related trace information swap.expired 1.0.1 ~ Internal command Manual call has no effect swap.scanexpire 1.0.1 ~ Internal command Manual call has no effect swap.mutexop 1.0.1 ~ Internal command Manual call has no effect -
GTID Related commands
Command Supported version Grammer Description Remark gtid 1.0.0 ~ Internal command gtidx 1.0.5 1) GTIDX [<arg> [value] [opt] ...]. Subcommands are:
2) LIST [<uuid>]
3) List gtid or uuid(if specified) gaps.
4) STAT [<uuid>]
5) Show gtid or uuid(if specified) stat.
6) REMOVE <uuid>
7) Return uuid.
8) HELP
9) Prints this help.gtidx function auxiliary command
-
-
ROR Supported Configurations
-
Redis Native Configuration ROR inherits all the native configurations of Redis, and only a few configuration semantics have changed. The specific list is as follows:
Features redis ROR production default (domestic) ROR production default (overseas) ror maxmemory The maximum amount of memory (i.e. the maximum capacity of keyspace) 2.5G 2.2G total memory:
256MB rocksdb writebuffer (max_write_buffer_num * write_buffer_size *3 (CF can consider setting different write_buffer*)) + blockcache
maxmemory: 2G settingThe maximum amount of memory, but the maximum capacity of the keyspace can exceed maxmemory (the excess is persisted to SSD). The maximum keyspace capacity of ROR is specified by swap-max-db-size. sanitize-dump-payload yes/no/client no no ror loads directly from rdb to rocksdb, skipping deep inspection. -
swap Related Configuration
Option name ROR version Value range Default value ROR production default (domestic) ROR production default (overseas) Can it be modified dynamically trocks option name trocks domestic trocks overseas Description swap-mode 1.0.1 disk, memory memory disk disk No ~ ~ ~ swap mode, disk: disk mode, enable hot and cold separation; memory: memory mode, the same as redis. swap-threads 1.0.1 4~64 4 4 4 No workers 4 4 The number of swap threads. swap-max-db-size 1.0.1 0~LLONG_MAX 0 40G 24G Yes max-db-size 40(G) 40(G) Maximum disk usage. swap-inprogress-memory-slowdown 1.0.1 0~LLONG_MAX 256mb 256mb 256mb Yes ~ ~ ~ When the memory occupied by the key and value being swapped exceeds swap-memory-slowdown, then ror Client write speed will begin to be throttled. swap-inprogress-memory-stop 1.0.1 0~LLONG_MAX 512mb 512mb 512mb Yes ~ ~ ~ When the memory occupied by the key and value being swapped exceeds swap-memory-slowdown, then ror Will start limiting client write speeds while pausing ejection of hot keys to disk. swap-maxmmeory-oom-percentage 1.0.1 100~INT_MAX 200 200 200 Yes ~ ~ ~ When the current used_memory of redis far exceeds maxmemory, ror will reject all write commands. swap-evict-step-max-subkeys 1.0.1 0~LLONG_MAX 1024 1024 1024 Yes ~ ~ ~ The maximum number of subkeys for a single evict (hot or cold) of a single key. swap-evict-step-max-memory 1.0.1 0~LLONG_MAX 1mb 1mb 1mb Yes ~ ~ ~ The maximum memory for a single evict (hot or cold) of a single key. swap-repl-max-rocksdb-read-bps 1.0.1 0~LLONG_MAX 0 100mb 100mb Yes max-replication-mb 100 100 When saving RDB in full copy, bgsave, etc. The maximum rate at which disk files can be read (i.e. the maximum rate at which iterates rocksdb data). swap-rocksdb-stats-collect-interval-ms 1.0.1 1~INT_MAX 2000 2000 2000 Yes ~ ~ ~ The interval for updating rocksdb stats cache. swap-evict-inprogress-limit 1.0.1 4~INT_MAX 64 64 64 Yes ~ ~ ~ The hard limit of the maximum concurrency of evict (hot and cold). The maximum concurrency of evict (hot and cold) is positively related to used_memory-maxmemory. The more memory exceeds the limit, the higher the concurrency of evict, but the concurrency of evict will not exceed swap-evict-inprogress-limit. swap-evict-inprogress-growth-rate 1.0.1 1~INT_MAX 10mb 10mb 10mb Yes ~ ~ ~ evict (hot and cold) concurrency growth rate, min (used_memory-maxmemory / swap-evict-inprogress-growth-rate, swap-evict-inprogress-limit) is the current evict concurrency limit swap-debug-trace-latency 1.0.1 yes,no no no no yes ~ ~ ~ Whether to enable the time statistics function of each stage of swap. swap-debug-evict-keys 1.0.1 -1~INT_MAX 0 0 0 Yes ~ ~ ~ The maximum number of evict (hot and cold) keys after the command ends, -1 means evict All hot keys. swap-debug-rio-delay-micro 1.0.1 -1~INT_MAX 0 0 0 Yes ~ ~ ~ The time (microseconds) for each RIO sleep. swap-debug-swapout-notify-delay-micro 1.0.1 -1~INT_MAX 0 0 0 Yes ~ ~ ~ The delay time (microseconds) for each notify. swap-debug-before-exec-swap-delay-micro 1.0.1 0~INT_MAX 0 0 0 Yes ~ ~ ~ The delay in microseconds before exec. swap-debug-init-rocksdb-delay-micro 1.0.1 0~INT_MAX 0 0 0 Yes ~ ~ ~ Open rocksdb delay time in microseconds. swap-perflog-sample-ratio 1.0.4 0~100 0 0 0 yes profiling-sample-ratio 0 0 perflog sampling rate swap-perflog-max-len 1.0.4 0~LONG_MAX 128 128 128 Yes profiling-sample-record-max-len 256 256 The maximum number of perflog records swap-perflog-log-slower-than-us 1.0.4 0~LONG_MAX 10000 10000 10000 Yes profiling-sample-record-threshold-ms 100 100 The threshold for perflog to retain samples, default When the swap operation takes more than 10ms, it will be recorded in perlog rocksdb.data.compression 1.0.3 no, snappy, zlib,optimized_for_compaction snappy snappy snappy yes rocksdb.compression snappy snappy Whether rocksdb enables compression feature. rocksdb.meta.compression 1.0.3 no, snappy, zlib,optimized_for_compaction snappy ~ ~ yes ~ ~ ~ rocksdb.meta.block_cache_size 1.0.1 0~ULLONG_MAX 512mb 512mb ? No rocksdb.metadata_block_cache_size 128 128 meta CF blockcache capacity rocksdb.data.block_cache_size 1.0.1 0~ULLONG_MAX 8mb 8mb 8mb(x2) No rocksdb.subkey_block_cache_size 128 128 data CF blockcache capacity rocksdb.max_open_files 1.0.1 -1~INT_MAX -1 -1 -1 No rocksdb.max_open_files 2048 2048 rocksdb max_open_files parameter value rocksdb.data.write_buffer_size 1.0.3 0~ULLONG_MAX 64mb 64mb ?(x3) Yes rocksdb.write_buffer_size 128 (actual 64) 16 (actual 64) rocksdb write_buffer_size parameter value rocksdb.meta.write_buffer_size 1.0.3 0~ULLONG_MAX 64mb ~ ~ Yes ~ ~ ~ rocksdb write_buffer_size parameter value rocksdb.data.target_file_size_base 1.0.3 0~ULLONG_MAX 32mb 32mb ? Yes rocksdb.target_file_size_base 128 16 (actual 128) rocksdb target_file_size_base parameter value rocksdb.meta.target_file_size_base 1.0.3 0~ULLONG_MAX 32mb ~ ~ Yes ~ ~ ~ rocksdb target_file_size_base parameter value rocksdb.data.max_write_buffer_number 1.0.3 1~256 3 3 4 Yes rocksdb.max_write_buffer_number 4 4 rocksdb max_write_buffer_number parameter value rocksdb.meta.max_write_buffer_number 1.0.3 1~256 3 ~ ~ Yes ~ ~ ~ rocksdb max_write_buffer_number parameter value rocksdb.max_background_compactions 1.0.1 1~64 4 4 4 No rocksdb.max_background_compactions 4 4 rocksdb max_background_compactions parameter value rocksdb.max_background_flushes 1.0.1 -1~64 -1 -1 -1 No rocksdb.max_background_flushes 4 4 rocksdb max_background_flushes parameter value rocksdb.max_subcompactions 1.0.1 1~64 1 1 1 No rocksdb.max_sub_compactions 4 4 rocksdb max_subcompactions parameter value rocksdb.data.block_size 1.0.3 512~INT_MAX 8192 8192 8192 No rocksdb.block_size 4096 4096 rocksdb block_size parameter value rocksdb.meta.block_size 1.0.3 512~INT_MAX 8192 ~ ~ No ~ ~ ~ rocksdb block_size parameter value rocksdb.cache_index_and_filter_blocks 1.0.3 yes no no no no no rocksdb.cache_index_and_filter_blocks no no rocksdb.meta.cache_index_and_filter_blocks 1.0.3 yes no no ~ ~ no ~ ~ ~ rocksdb.enable_pipelined_write 1.0.1 yes|no no no no No rocksdb.enable_pipelined_write no no Whether to enable the pipelined write function of RocksDB. rocksdb.data.level0_slowdown_writes_trigger 1.0.3 1~INT_MAX 20 20 20 Yes rocksdb.level0_slowdown_writes_trigger 24 24 Threshold to trigger Level 0 write slowdown. rocksdb.meta.level0_slowdown_writes_trigger 1.0.3 1~INT_MAX 20 ~ ~ Yes ~ ~ ~ Threshold to trigger Level 0 write slowdown. rocksdb.data.disable_auto_compactions 1.0.3 yes|no no no no Yes rocksdb.disable_auto_compactions no no Whether to disable automatic compaction. rocksdb.meta.disable_auto_compactions 1.0.3 yes|no no ~ ~ Yes rocksdb.disable_auto_compactions no no Whether to disable automatic compaction. rocksdb.ratelimiter.rate_per_sec 1.0.3 0~ULLONG_MAX 512mb ~ ~ No rocksdb.ratelimiter.rate_per_sec no no Maximum IO rate for compaction and flushing. If set, the delayed write rate (maximum DB write rate after slowing down) will be set to the same value. rocksdb.bytes_per_sync 1.0.3 0~ULLONG_MAX 1mb ~ ~ No ~ no no Perform fsync every bytes_per_sync during SST file creation. rocksdb.data.max_bytes_for_level_base 1.0.3 0~ULLONG_MAX 256mb ~ ~ Yes ~ ~ ~ Maximum file size for Level 1. rocksdb.meta.max_bytes_for_level_base 1.0.3 0~ULLONG_MAX 256mb ~ ~ Yes ~ ~ ~ Maximum file size for Level 1. rocksdb.data.max_bytes_for_level_multiplier 1.0.5 1~INT_MAX 10 ~ ~ Yes ~ ~ ~ Multiplier for the maximum file size for each level. Ln+1 = Ln * multiplier. rocksdb.meta.max_bytes_for_level_multiplier 1.0.5 1~INT_MAX 10 ~ ~ Yes ~ ~ ~ Multiplier for the maximum file size for each level. Ln+1 = Ln * multiplier. rocksdb.data.compaction_dynamic_level_bytes 1.0.5 yes|no no ~ ~ No ~ ~ ~ Dynamic level capacity mode, dynamically calculate the maximum capacity of each level based on the data set size. Optimization space can be extended to 1.111111. rocksdb.meta.compaction_dynamic_level_bytes 1.0.5 yes|no no ~ ~ No ~ ~ ~ Dynamic level capacity mode, dynamically calculate the maximum capacity of each level based on the data set size. Optimization space can be extended to 1.111111. rocksdb.data.suggest_compact_deletion_percentage 1.0.5 0~100 0.95 ~ ~ No ~ ~ ~ Mark SST files when RocksDB is idle if the percentage of deletion marks in the SST file is greater than this configuration and perform compaction. 0 disables this feature. rocksdb.meta.suggest_compact_deletion_percentage 1.0.5 0~100 0.95 ~ ~ No ~ ~ ~ Mark SST files when RocksDB is idle if the percentage of deletion marks in the SST file is greater than this configuration and perform compaction. 0 disables this feature. rocksdb.data.periodic_compaction_seconds 1.0.5 0~ULLONG_MAX 86400 ~ ~ Yes ~ ~ ~ Automatically perform compaction at certain intervals after the SST file is generated. This compaction does not push SST files to lower levels, it is only used for triggering swap-scan-session-bits 1.0.5 1~16 7 No The maximum number of scan sessions is (1<<swap-scan-session-bits) swap-scan-session-max-idle-seconds 1.0.5 1~INT_MAX 60 Scan session maximum idle time (session exceeding this idle time can be used again) ctrip-monitor-port 1.0.7 0~65535 6380 Monitoring port swap-absent-cache-enabled 1.0.7 yes no yes Whether to start absent cache. (Absent cache caches recently accessed non-existent keys in LRU mode, which is used to solve the delay problem of hotspot access to non-existent keys) swap-absent-cache-capacity 1.0.7 1~LLONG_MAX 65536 The capacity of the absent cache, indicating the number of keys that can be cached. swap-batch-limit 1.1.0 "IN 16 1mb OUT 16 1mb DEL 16 1mb" batch size swap-cuckoo-filter-enabled 1.1.0 yes no yes cuckoo filter switch swap-cuckoo-filter-bit-per-key 1.1.0 8 12 16 32 8 The bits occupied by each key in cuckoo filter swap-cuckoo-filter-estimated-keys 1.1.0 1~LLONG_MAX 32000000 Estimated number of keys swap-absent-cache-include-subkey 1.1.0 yes|no yes yes ~ ~ ~ ~ ~ absent cache includes subkey switch rocksdb.max_background_jobs 1.1.0 -1~64 4 No ~ ~ ~ ~ ~ rocksdb maximum number of background jobs swap-ratelimit-maxmemory-percentage 1.1.1 100~INT_MAX 200 yes ~ -ratelimit-policydecision swap-ratelimit-policy 1.1.1 pause|reject_oom|reject_all|disabled pause Yes ~ ~ ~ ~ ~ ror's rate limiting behavior on the client, including 4 types: pause, reject_oom, reject_all, disabled swap-ratelimit-maxmemory-pause-growth-rate 1.1.1 1~INT_MAX 20mb Yes ~ ~ ~ ~ ~ When ror uses the pause policy to limit the speed of the client, the speed limit time growth rate swap-compaction-filter-skip-level ~ -1~INT_MAX 0 Yes ~ ~ ~ ~ ~ The level skipped by compaction filter swap-compaction-filter-disable-until 1.1.1 0~LLONG_MAX 0 Yes ~ ~ ~ ~ ~ Temporarily turn off compaction-filter until the specified timestamp, in unix timestamps (seconds level) swap-persist-enabled 1.2.0 yes|no no no ~ ~ ~ ~ ~ Turn on persistence function swap-ratelimit-persist-lag 1.2.0 0~INT_MAX 60 Yes ~ ~ ~ ~ ~ Persistence delay rate limit threshold, after exceeding swap-ratelimit-persist-lag, it will be based on Selected swap-ratelimit-policy for rate limiting swap-ratelimit-persist-pause-growth-rate 1.2.0 0~INT_MAX 10 is ~ ~ ~ ~ ~ lag per growth swap-ratelimit-persist-pause-growth-rate, The speed limit time is increased by 1ms. rocksdb.WAL_ttl_seconds 1.2.0 0~INT_MAX 18000 No ~ ~ ~ ~ ~ rocksdb WAL maximum retention time rocksdb.WAL_size_limit_MB 1.2.0 0~INT_MAX 16384 No ~ ~ ~ ~ ~ rocksdb WAL maximum reserved capacity rocksdb.max_total_wal_size 1.2.0 0~ULONGLONG_MAX 536870912 No ~ ~ ~ ~ ~ When the wal capacity exceeds this value, the cf flush that depends on the wal is triggered. swap-debug-rdb-key-save-delay-micro 1.2.0 -1~INT_MAX 0 Yes ~ ~ ~ ~ ~ rdbsave cold key delay time rocksdb.data.suggest_compact_sliding_window_size 1.2.1 0~ULONGLONG_MAX 100000 No ~ ~ rocksdb.suggest_compact_sliding_window_size ~ ~ suggest compaction trigger sliding window size rocksdb.meta.suggest_compact_sliding_window_size 1.2.1 0~ULONGLONG_MAX 100000 No ~ ~ ~ ~ ~ suggest compaction trigger sliding window size rocksdb.data.suggest_compact_num_dels_trigger 1.2.1 0~ULONGLONG_MAX 80000 No ~ ~ rocksdb.suggest_compact_num_dels_trigger ~ ~ The number of deletes in the sliding window that triggers suggestion compaction rocksdb.meta.suggest_compact_num_dels_trigger 1.2.1 0~ULONGLONG_MAX 80000 No ~ ~ ~ ~ ~ The number of deletes in the sliding window that triggers suggestion compaction swap-flush-meta-deletes-percentage 1.2.1 0~100 40 Yes ~ ~ ~ ~ ~ The delete proportion that triggers METACF memtable flush (swap-flush-meta-deletes-percentage && swap-flush-meta-deletes-num will trigger flush only when they exceed the threshold) swap-flush-meta-deletes-num 1.2.1 1~LLONG_MAX 200000 Yes ~ ~ ~ ~ ~ The delete proportion that triggers METACF memtable flush (swap-flush-meta-deletes-percentage && swap-flush-meta-deletes-num will trigger flush only when they exceed the threshold) rocksdb.data.enable_blob_files 1.2.3 yes|no no Yes ~ ~ rocksdb.enable_blob_files ~ ~ Whether to enable blobdb rocksdb.meta.enable_blob_files 1.2.3 yes|no no Yes ~ ~ ~ ~ ~ Whether to enable blobdb rocksdb.data.enable_blob_garbage_collection 1.2.3 yes|no no yes ~ ~ rocksdb.enable_blob_garbage_collection ~ ~ Whether to enable blobdb force GC rocksdb.meta.enable_blob_garbage_collection 1.2.3 yes|no no yes ~ ~ ~ ~ ~ Whether to enable blobdb force GC rocksdb.data.blob_garbage_collection_age_cutoff_percentage 1.2.3 0~100 5 Yes ~ ~ ~ ~ ~ blobdb forces GC selection range (by default, only the oldest 5% of blobs will be selected by GC )
-
-
ROR INFO information
-
Redis native INFO ROR inherits all native info indicators of Redis, and its meaning remains the same as Redis.
-
SWAP Related INFO
section version field description Cpu 1.1.0 swap_main_thread_cpu_usage ror main thread cpu usage 1.1.0 swap_swap_threads_cpu_usage ror swap thread cpu usage 1.1.0 swap_other_threads_cpu_usage ror other thread cpu usage Memory 1.0.1 swap_mem_rocksdb rocksdb amount of memory occupied 1.0.1 swap_rectified_frag_ratio eliminate the memory fragmentation rate of rocksdb occupied memory 1.0.1 swap_rectified_frag_bytes eliminate the amount of memory fragmentation occupied by rocksdb Keyspace 1.0.1 Keyspace:"db%d:keys=%lld,evicts=%lld,metas=%lld,expires=%lld,avg_ttl=%lld\r\n" evicts:number of cold keys metas:number of objectMeta
-
-
ROR inherits all the native configurations of Redis, and only a few configuration semantics have changed. The specific list is as follows:
Redis configuration for semantic changes Characteristics redis semantics ror production default (domestic) ror production default (overseas) ror semantics sanitize-dump-payload yes no client three options no no ror is loaded directly from rdb to rocksdb, skipping depth inspection. maxmemory Maximum amount of memory (i.e. maximum capacity of keyspace) 2.5G 2.2G total memory:
256MB rocksdb
writebuffer(max_write_buffer_num * write_buffer_size *3 (CF can consider setting different write_buffer*))+blockcache
maxmemory: 2G settingsThe maximum amount of memory, but the maximum capacity of the keyspace can exceed maxmemory (the excess is persisted to SSD). The maximum keyspace capacity of ROR is specified by swap-max-db-size. -
SWAP related configuration
Option name ROR version Value range Default value ROR production default (domestic) ROR production default (overseas) Can it be modified dynamically Description ctrip-monitor-port 1.0.7 0~65535 6380 No Monitoring port rocksdb.bytes_per_sync 1.0.3 0~ULLONG_MAX 1mb No During the creation of rocksdb SST files, fsync is executed every bytes_per_sync rocksdb.cache_index_and_filter_blocks 1.0.3 yes|no no no no No rocksdb cache_index_and_filter_blocks parameter value rocksdb.data.blob_file_size 1.2.3 0 ~ ULLONG_MAX 256mb Yes blobdb blob file size rocksdb.data.blob_garbage_collection_age_cutoff_percentage 1.2.3 0~100 5 Yes blobdb forces GC selection range (by default, only the oldest 5% of blobs will be selected by GC) rocksdb.data.blob_garbage_collection_force_threshold_percentage 1.2.3 0~100 90 Yes The threshold for blobdb to force GC (by default, only blobs with more than 90% garbage data will force GC) rocksdb.data.block_cache_size 1.0.1 0~ULLONG_MAX 8mb 8mb 8mb(x2) No data CF blockcache capacity