• Google Group
• Slack Channel

Kvrocks is a distributed key value NoSQL database that uses RocksDB as storage engine and is compatible with Redis protocol. Kvrocks intends to decrease the cost of memory and increase the capability while compared to Redis. The design of replication and storage was inspired by rocksplicator and blackwidow.

Kvrocks has the following key features:

• Redis protocol, user can use redis client to visit the kvrocks
• Namespace, similar to redis db but use token per namespace
• Replication, async replication using binlog like MySQL
• High Available, supports redis sentinel to failover when master or slave was failed
• Cluster mode, centralized management but compatible with Redis cluster client access

Thanks for designers @田凌宇 and @范世丽 contribute the kvrocks logo for us.

Who uses kvrocks

Tickets a pull reqeust to let us known that you're using kvrocks and add your logo to README

Building kvrocks

requirements

• g++ (required by c++11, version >= 4.8)
• autoconf automake libtool snappy

Build

NOTE: You should install the snappy first:

# Centos/Redhat
sudo yum install -y epel-release && sudo yum install -y git gcc gcc-c++ make snappy snappy-devel autoconf automake libtool which gtest gtest-devel

# Ubuntu/Debian
sudo apt update
sudo apt-get install gcc g++ make libsnappy-dev autoconf automake libtool googletest libgtest-dev

# MACOSX
brew install autoconf automake libtool snappy googletest

It is as simple as:

$ git clone --recursive https://github.com/apache/incubator-kvrocks.git
$ cd kvrocks
$ make -j4

Running kvrocks

$ ./src/kvrocks -c kvrocks.conf

Running test cases

NOTE: You should install the googletest first

make test

Supported platforms

• centos 6/7
• ubuntu
• macosx

Try kvrocks using Docker

$ docker run -it -p 6666:6666 kvrocks/kvrocks
$ redis-cli -p 6666

127.0.0.1:6666> get a
(nil)

Namespace

namespace was used to isolate data between users. unlike all the redis databases can be visited by requirepass, we use one token per namespace. requirepass was regraded as admin token, only admin token allows to access the namespace command, as well as some commands like config, slaveof, bgsave, etc…

# add token
127.0.0.1:6666>  namespace add ns1 my_token
OK

# update token
127.0.0.1:6666> namespace set ns1 new_token
OK

# list namespace
127.0.0.1:6666> namespace get *
1) "ns1"
2) "new_token"
3) "__namespace"
4) "foobared"

# delete namespace
127.0.0.1:6666> namespace del ns1
OK

Cluster

Kvrocks implements a proxyless centralized cluster solution but its accessing method is completely compatible with the Redis cluster client. You can use Redis cluster SDKs to access the kvrocks cluster. More details, please see: Kvrocks Cluster Introduction

DOCs

• supported commands
• design complex kv on rocksdb
• replication design

For more informations, can see: Kvrocks Wiki

Tools

• Export the Kvrocks monitor metrics, please use kvrocks_exporter
• Migrate from redis to kvrocks, use redis-migrate-tool which was developed by vipshop
• Migrate from kvrocks to redis. use kvrocks2redis in build dir

Performance

Hardware

• CPU: 48 cores Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
• Memory: 32 GiB
• NET: Intel Corporation I350 Gigabit Network Connection
• DISK: 2TB NVMe Intel SSD DC P4600

Benchmark Client: multi-thread redis-benchmark(unstable branch)

1. Commands QPS

kvrocks: workers = 16, benchmark: 8 threads/ 512 conns / 128 payload

latency: 99.9% < 10ms

2. QPS on different payloads

kvrocks: workers = 16, benchmark: 8 threads/ 512 conns

latency: 99.9% < 10ms

3. QPS on different workers

kvrocks: workers = 16, benchmark: 8 threads/ 512 conns / 128 payload

latency: 99.9% < 10ms

License

Kvrocks is under the Apache License Version 2.0. See the LICENSE file for details.

微信公众号