Friday, December 4, 2015

Remove Shards From MongbDB 
Sharded Cluster


To remove a shard you must ensure the shard’s data is migrated to the remaining shards in the cluster.

In my env, I have a sharded and replicated cluster where each shard has two members i.e. primary and seconday. 


To remove a shard, first connect to one of the cluster’s mongos instances using mongo shell. 

[root@xxxx  mongodb]# mongo --host xxxx--port 20001
MongoDB shell version: 3.0.7
connecting to: lpdosput00249:20001/test
Server has startup warnings:
2015-11-19T09:59:29.367-0700 I CONTROL
mongos> db.adminCommand( { listShards: 1 } )
{
        "shards" : [
                {
                        "_id" : "shard0000",
                        "host" : "rs0/10.20.176.93:30001,10.20.176.93:31001"
                },
                {
                        "_id" : "shard0001",
                        "host" : "rs1/10.20.176.93:30002,10.20.176.93:31002"
                },
                {
                        "_id" : "shard0002",
                        "host" : "rs2/10.20.176.93:30003,10.20.176.93:31003"
                }
        ],
        "ok" : 1
}

To successfully migrate data from a shard, the balancer process must be enabled. 
Check the balancer state using the sh.getBalancerState() helper in the mongo shell

mongos> sh.getBalancerState()
true

Now, one needs to determine the shard to remove. Make sure you pick up the right name, otherwise you will end removing wrong shard. This operation is IO intensive and will take time proportional to the volume of data to be moved. 

mongos> use admin
switched to db admin
mongos> db.runCommand( { removeShard: "shard0000" } )
{
        "msg" : "draining started successfully",
        "state" : "started",
        "shard" : "shard0000",
        "note" : "you need to drop or movePrimary these databases",
        "dbsToMove" : [
                "testdb"
        ],
        "ok" : 1
}
mongos>

As you can see the message that the draining of data is successfully started. This begins “draining” chunks from the shard you are removing to other shards in the cluster.


mongos> sh.status()
--- Sharding Status ---
  sharding version: {
        "_id" : 1,
        "minCompatibleVersion" : 5,
        "currentVersion" : 6,
        "clusterId" : ObjectId("564659a6ecf5ebdaa8b6fec9")
}
  shards:
        {  "_id" : "shard0000",  "host" : "rs0/10.20.176.93:30001,10.20.176.93:31001",  "draining" : true }
        {  "_id" : "shard0001",  "host" : "rs1/10.20.176.93:30002,10.20.176.93:31002" }
        {  "_id" : "shard0002",  "host" : "rs2/10.20.176.93:30003,10.20.176.93:31003" }
  balancer:
        Currently enabled:  yes
        Currently running:  no
        Failed balancer rounds in last 5 attempts:  3
        Last reported error:  ReplicaSetMonitor no master found for set: rs0
        Time of Reported error:  Wed Nov 18 2015 16:02:28 GMT-0700 (MST)
        Migration Results for the last 24 hours:
                No recent migrations
  databases:
        {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
        {  "_id" : "testdb",  "partitioned" : true,  "primary" : "shard0000" }
        {  "_id" : "test_db",  "partitioned" : true,  "primary" : "shard0001" }
        {  "_id" : "apiv3",  "partitioned" : false,  "primary" : "shard0001" }
        {  "_id" : "mmsdbpings",  "partitioned" : false,  "primary" : "shard0001" }
        {  "_id" : "test",  "partitioned" : false,  "primary" : "shard0001" }


mongos>

To check the progress of the migration at any stage in the process, run removeShard from the admin database again. 

mongos> db.runCommand( { removeShard: "shard0000" } )
{
        "msg" : "draining ongoing",
        "state" : "ongoing",
        "remaining" : {
                "chunks" : NumberLong(0),
                "dbs" : NumberLong(1)
        },
        "note" : "you need to drop or movePrimary these databases",
        "dbsToMove" : [
                "testdb"
        ],
        "ok" : 1
}


Now try to move primary shard for the testdb database. 

mongos> db.runCommand( { movePrimary: "testdb", to: "shard0001" })
{
        "primary " : "shard0001:rs1/10.20.176.93:30002,10.20.176.93:31002",
        "ok" : 1
}

check the status of the drain operation...

mongos> db.runCommand( { removeShard: "shard0000" } )
{
        "msg" : "removeshard completed successfully",
        "state" : "completed",
        "shard" : "shard0000",
        "ok" : 1
}

Now if you check the status of the sharded cluster you will see only two shards available. 

mongos> sh.status()
--- Sharding Status ---
  sharding version: {
        "_id" : 1,
        "minCompatibleVersion" : 5,
        "currentVersion" : 6,
        "clusterId" : ObjectId("564659a6ecf5ebdaa8b6fec9")
}
  shards:
        {  "_id" : "shard0001",  "host" : "rs1/10.20.176.93:30002,10.20.176.93:31002" }
        {  "_id" : "shard0002",  "host" : "rs2/10.20.176.93:30003,10.20.176.93:31003" }
  balancer:
        Currently enabled:  yes
        Currently running:  no
        Failed balancer rounds in last 5 attempts:  0
        Migration Results for the last 24 hours:
                No recent migrations
  databases:
        {  "_id" : "admin",  "partitioned" : false,  "primary" : "config" }
        {  "_id" : "testdb",  "partitioned" : true,  "primary" : "shard0001" }
        {  "_id" : "test_db",  "partitioned" : true,  "primary" : "shard0001" }
        {  "_id" : "apiv3",  "partitioned" : false,  "primary" : "shard0001" }
        {  "_id" : "mmsdbpings",  "partitioned" : false,  "primary" : "shard0001" }
        {  "_id" : "test",  "partitioned" : false,  "primary" : "shard0001" }

Hope this will help you to remove the replicated shards from your cluster. For non replicated shards the process will remain same. 



No comments:

Post a Comment