Support cluster mode

[ShooterIT]

resolves #219

Deploy Kvrocks Cluster

In our cluster solution, you should firstly deploy some kvrocks instances and set replication relationship, and
Kvrocks doesn't support to communicate with each others, so we need to tell every node the cluster topology.
The command is CLUSTER SETNODES

• CLUSTERX SETNODES $ALL_NODES_INFO $VERSION $FORCE
$ALL_NODES_INFO format: $node_id $ip $port $role $master_node_id $slot_range. $role: master or slave, $master_node_id: master_node_id if it is slave, otherwise, this value is -. slave needn't set $slot_range.
  Example:

  07c37dfeb235213a872192d90877d0cd55635b91 127.0.0.1 30004 slave e7d1eecce10fd6bb5eb35b9f99a514335d9ba9ca
  67ed2db8d677e59ec4a4cefb06858cf2a1a89fa1 127.0.0.1 30002 master - 5461-10922
  

  $VERSION means current cluster topology version that is stored or generated by management component such as config server or meta-server.

Kvrocks can know itself node id by compared with itself ip and port in received all nodes info, but you maybe bind kvrorks to 0.0.0.0, so it is necessary to set NODEID explicitly to let it know itself. As the same with redis cluster, node id is 40 characters length.

• CLUSTERX SETNODEID $NODE_ID

Moreover, cluster doesn't handle failover and migration, you should set cluster topology again after failover, of course, you should do that if you add some new nodes into cluster.

Access Kvrocks Cluster

After setting cluster topology, we have tested, it adopt redis-cli, redis-benchmark, python redis cluster sdk, and redis-cluster-proxy. So you may doesn't need to change your redis cluster request code to access kvrocks cluster.

Have a try

In tool/try_cluster, you can change BIN_PATH of kvrocks bin path, and then
./try_cluster.sh start to create a small cluster that has 6 nodes(3 maters and 3 replicas)

then run ./try_cluster.sh watch to observe cluster nodes, you also can do some tests by
redis-cli, redis-benchmark, redis cluster sdk and redis-cluster-proxy for your fun.

finally you can run ./try_cluster.sh stop to stop all nodes and clean up directory.

Supplements

Use CLUSTERX command to set cluster topology instead of CLUSTER #324

[git-hulk]

LGTM, the cluster topology and version would store on local storage?

[ShooterIT]

@git-hulk That is a good question, I still consider that, we must should know the nodes may keep old cluster topology if restarted. That may cause terrible things if the IP and port are reused.

[git-hulk]

yup, use ip:port as the unique identifier would cause problems, the node id would be better if we also persist it.