Mongodb adds and removes instances of Shard Server

1. Add a new configuration file and start the mongod instance

 

#14

mkdir -p /data/mongodb/shard311

./mongod -f /opt/mongodb-linux-x86_64-2.2.0/conf/mongod_3.conf

 

#16

mkdir -p /data/mongodb/shard32

./mongod -f /opt/mongodb-linux-x86_64-2.2.0/conf/mongod_3.conf

 

#23

mkdir -p /data/mongodb/shard33

./mongod -f /opt/mongodb-linux-x86_64-2.2.0/conf/mongod_3.conf

Step 2: initialize the set cluster

/opt/mongodb-linux-x86_64-2.2.0/bin/mongo -port 10003

config = {_id: 'shard3', members: [

         {_id: 0, host: '192.168.1.14:10003', priority:1},

         {_id: 1, host: '192.168.1.16:10003'},

         {_id: 2, host: '192.168.1.23:10003'}]};

rs.initiate(config);

3. Increase shard

/opt/mongodb-linux-x86_64-2.2.0/bin/mongo 192.168.1.14:10000/admin

db.runCommand( {

    addshard : "shard3/192.168.1.14:10003,192.168.1.16:10003,192.168.1.23:10003",

    name:"shard3",

    maxsize:20480,

    allowLocal:true } );

4. Remove the shard

db.runCommand({"removeshard" : "shard3/192.168.1.14:10003,192.168.1.16:10003,192.168.1.23:10003"});

Note:

The three servers clock are out of sync, which leads to the problem that cannot be shard. Today, I have encountered it again. It will be ok after synchronization. Is this a bit of a frequent problem? Should clock of 3 servers be synchronized at a fixed time every day?
It will take a long time to remove shard, and printShardingStatus() will display the status "draining" : true.