CKV Edition (Cluster Architecture)

Last updated: 2020-06-12 14:26:28

    TencentDB for Redis CKV Edition (cluster architecture) provides dual-replica cluster edition instances, which break the single-thread bottleneck to meet your business needs for large capacity or high performance. It is compatible with Redis 3.2 protocols and commands and supports up to 128 shards of 12 GB–48 TB capacity.

    Features

    • Service reliability
      With a dual-server master/slave architecture, the master and slave nodes reside on different physical machines with the master node providing external access. You can perform data CRUD using the Redis command line or client. The slave node provides data backup and high availability (HA). In case that the master node fails, the proprietary HA system will automatically perform master-slave switch to ensure smooth operation of the business.
    • Data reliability
      The data persistence feature is enabled by default, so all data will be stored in disks. Data backup is supported; therefore, you can roll back or clone instances for backup sets to effectively cope with data misoperations and other issues.
    • Lower latency
      CKV uses a high-performance network platform and a proxy-free architecture, which significantly reduce the access latency and network latency by up to 60% in high-load scenarios.
    • Read-only slave
      CKV Edition (cluster architecture) can greatly improve the read performance by 40% on average by enabling slaves. The read-only slave feature is not enabled by default. Currently, you can submit a ticket for application. Due to the replication delay between the CKV master node and slave node, after the real-only slave feature is enabled, some legacy data may be read; therefore, please confirm whether your business can accept slight data inconsistency before enabling this feature.
    • Smooth upgrade
      CKV Edition (cluster architecture) uses a unique scheme to ensure that version upgrade is imperceptible to the business, thereby ensuring maximum service availability.

    Use Cases

    • Single instance for high data volume
      CKV Edition (cluster architecture) uses a distributed architecture, making it suitable for storing high volumes of data in one single instance. Its capacity can exceed the upper limit of 384 GB of CKV Standard Edition.
    • High QPS and concurrence requirements
      CKV Edition (cluster architecture) uses a distributed architecture where reads and writes are spread across multiple nodes. Under the condition that the keys are evenly distributed, its QPS can increase linearly with the number of nodes. At present, it supports a maximum of 128 shards and 10 million QPS.
    • Insensitivity for protocol support
      CKV Edition (cluster architecture) supports slightly less protocols than the open-source editions.

    Connection Sample

    CKV Edition (cluster architecture) only supports the password format of instance ID:password. For example, if your instance ID is crs-bkuza6i3 and the password is abcd1234, then the connection command should be redis-cli -h IP -p port -a crs-bkuza6i3:abcd1234.

    Use Limits

    -The minimum unit of pttl setting display in the CKV engine is second, which is different from Redis Community Edition.

    • Currently, string-type keys are supported, and a value can be up to 32 MB, which are different from Redis Community Edition.
    • Except batch operations of MSET and MGET, other batch operations require the batch keys to be in the same slot; otherwise, an error will be displayed stating CROSSSLOT Keys in request don't hash to the same slot.
    • After a shard is full, subscribe and psubscribe operations need to use a certain amount of memory; therefore, new subscriptions will be affected, but publish operations on subscribed channels will not.

    Special Notes

    • At present, the size of a single shard in CKV Edition (cluster architecture) is 4 GB by default; therefore, it is recommended that the value of a single key not exceed 4 GB.
    • CKV Edition (cluster architecture) currently provides monitoring at the cluster dimension.

    Compatibility

    Commands supported by CKV Edition (cluster architecture):

    connection group geo group hashes group hyperloglog group keys group lists group pub/sub group
    auth geoadd hdel pfadd del lindex psubscribe
    echo geohash hexists pfcount exists linsert pubsub
    ping geopos hget pfmerge expire llen publish
    quit geodist hgetall - expireat lpop punsubscribe
    select georadius hincrby - type lpush subscribe
    - georadiusbymember hincrbyfloat - ttl lpushx unsubscribe
    - - hkeys - persist lrange -
    - - hlen - pexpire lrem -
    - - hmget - pexpireat lset -
    - - hmset - pttl ltrim -
    - - hset - rename rpop -
    - - hsetnx - renamenx rpoplpush -
    - - hstrlen - sort rpush -
    - - hvals - - rpushx -
    - - hscan - - - -
    sets group sorted sets group strings group transactions group server group
    sadd zadd append discard command
    scard zcard bitcount exec dbsize
    sdiff zcount bitop multi -
    sdiffstore zincrby bitpos unwatch -
    sinter zinterstore decr watch -
    sinterstore zlexcount decrby - -
    sismember zrange get - -
    smembers zrangebylex getbit - -
    smove zrangebyscore getrange - -
    spop zrank getset - -
    srandmember zrem incr - -
    srem zremrangebylex incrby - -
    sscan zremrangebyrank incrbyfloat - -
    sunion zremrangebyscore mget - -
    sunionstore zrevrange mset - -
    - zrevrangebylex msetnx - -
    - zrevrangebyscore psetex - -
    - zrevrank set - -
    - zscan setbit - -
    - zscore setex - -
    - zunionstore setnx - -
    - - setrange - -
    - - strlen - -

    Commands not supported by CKV Edition (cluster architecture):

    cluster group connection group keys group lists group scripting group server group strings group
    cluster addslots swapdb touch blpop eval bgrewriteaof bitfield
    cluster count-failure-reports - restore brpop evalsha bgsave -
    cluster delslots - object brpoplpush script debug client kill -
    cluster failover - unlink - script exists client list -
    cluster forget - wait - script flush client getname -
    cluster meet - migrate - script kill client pause -
    cluster replicate - dump - script load client reply -
    cluster reset - scan - - client setname -
    cluster saveconfig - keys - - command count -
    cluster set-config-epoch - move - - command getkeys -
    cluster setslot - randomkey - - command info -
    cluster slaves - - - - config get -
    readonly - - - - config rewrite -
    readwrite - - - - config set -
    - - - - - config resetstat -
    - - - - - debug object -
    - - - - - debug segfault -
    - - - - - flushall -
    - - - - - flushdb -
    - - - - - lastsave -
    - - - - - monitor -
    - - - - - role -
    - - - - - save -
    - - - - - shutdown -
    - - - - - slaveof -
    - - - - - slowlog -
    - - - - - sync -
    - - - - - info -

    Was this page helpful?

    Was this page helpful?

    • Not at all
    • Not very helpful
    • Somewhat helpful
    • Very helpful
    • Extremely helpful
    Send Feedback
    Help