1. 云图网首页
  2. 技术专区
  3. 数据库

Sharding：Shard Cluster 删除 Shard 节点

• 数据库

上篇Blog学习了给 Shard Cluster 增加 shard 节点，接下来学习下删除节点。

一 删除 Shard 节点基本思路

  1. 移走要删除 Shard 节点的数据；
  2. 确保要删除的 shard 节点不是任何库的 Primary 节点，如果是，将这些库的 Primary 节点移到其它 shard 节点；
  3. 从 Shard Cluster 配置中删除 shard 节点信息。

二 删除 Shard 节点步骤

2.1 查看当前 Shard Cluster 状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
 
mongos> sh.status();  
--- Sharding Status ---  
 sharding version: { "_id" : 1, "version" : 3 }  
 shards:  
 { "_id" : "shard0000", "host" : "redhatB.example.com:5281" }  
 { "_id" : "shard0001", "host" : "redhatB.example.com:5282" }  
 { "_id" : "shard0002", "host" : "redhatB.example.com:5283" }  
 databases:  
 { "_id" : "admin", "partitioned" : false, "primary" : "config" }  
 { "_id" : "test", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "francs", "partitioned" : true, "primary" : "shard0000" }  
 francs.test_1 chunks:  
 shard0000 2  
 shard0001 2  
 shard0002 3  
 { "id" : { $minKey : 1 } } -->> { "id" : 1 } on : shard0000 Timestamp(5000, 0)  
 { "id" : 1 } -->> { "id" : 7282 } on : shard0001 Timestamp(5000, 1)  
 { "id" : 7282 } -->> { "id" : 15651 } on : shard0001 Timestamp(4000, 3)  
 { "id" : 15651 } -->> { "id" : 29868 } on : shard0002 Timestamp(4000, 4)  
 { "id" : 29868 } -->> { "id" : 37149 } on : shard0002 Timestamp(5000, 2)  
 { "id" : 37149 } -->> { "id" : 48037 } on : shard0002 Timestamp(5000, 3)  
 { "id" : 48037 } -->> { "id" : { $maxKey : 1 } } on : shard0000 Timestamp(4000, 1)  
 { "_id" : "records", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "fracns", "partitioned" : false, "primary" : "shard0002" }

mongos> db.test_1.count();  
100000

备注：当前有三个 shard 节点，其中 test_1 集合有 10000 个文档。目标删除 “redhatB.example.com:5283” 节点。

2.2 从目标 shard 中转移数据
使用 removeShard 命令就能将 shard 节点的数据转移到其它节点，前提是已开启balancer 进程。

2.2.1 查看 balancer 进程是否已开启

1
2
 
mongos> sh.getBalancerState();  
true

2.2.2 转移数据

1
2
3
4
5
6
7
8
9
10
11
12
13
 
mongos> use admin;  
switched to db admin

mongos> db.runCommand({removeshard:"redhatB.example.com:5283"});{  
 "msg" : "draining started successfully",  
 "state" : "started",  
 "shard" : "shard0002",  
 "note" : "you need to drop or movePrimary these databases",  
 "dbsToMove" : [  
 "fracns"  
 ],  
 "ok" : 1  
}

备注：这个步骤会将目标删除的 shard 节点数据转移到其它 shard 节点，数据量越大，耗时越长。

2.3 核查迁移状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
 
mongos> db.runCommand({removeshard:"redhatB.example.com:5283"});  
{  
 "msg" : "draining ongoing",  
 "state" : "ongoing",  
 "remaining" : {  
 "chunks" : NumberLong(0),  
 "dbs" : NumberLong(1)  
 },  
 "note" : "you need to drop or movePrimary these databases",  
 "dbsToMove" : [  
 "fracns"  
 ],  
 "ok" : 1  
}

备注： 提示所删除的 shard 节点是某个库的 Primary 节点，再根据 2.1 红色字体信息，发现 “redhatB.example.com:5283” 节点是 fracns 库的主节点。

2.4 转移 Unsharded Databases

1
2
3
4
5
 
mongos> use admin;  
switched to db admin

mongos> db.runCommand( { movePrimary: "fracns", to: "redhatB.example.com:5281" }) ;  
{ "primary " : "shard0000:redhatB.example.com:5281", "ok" : 1 }

备注：这个命令将 fracns 库的主节点转移到 “redhatB.example.com:5281” shard 节点。

2.5 查看 Shard Cluster 状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
 
mongos> sh.status();  
--- Sharding Status ---  
 sharding version: { "_id" : 1, "version" : 3 }  
 shards:  
 { "_id" : "shard0000", "host" : "redhatB.example.com:5281" }  
 { "_id" : "shard0001", "host" : "redhatB.example.com:5282" }  
 { "_id" : "shard0002", "draining" : true, "host" : "redhatB.example.com:5283" }  
 databases:  
 { "_id" : "admin", "partitioned" : false, "primary" : "config" }  
 { "_id" : "test", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "francs", "partitioned" : true, "primary" : "shard0000" }  
 francs.test_1 chunks:  
 shard0000 4  
 shard0001 3  
 { "id" : { $minKey : 1 } } -->> { "id" : 1 } on : shard0000 Timestamp(5000, 0)  
 { "id" : 1 } -->> { "id" : 7282 } on : shard0001 Timestamp(5000, 1)  
 { "id" : 7282 } -->> { "id" : 15651 } on : shard0001 Timestamp(4000, 3)  
 { "id" : 15651 } -->> { "id" : 29868 } on : shard0000 Timestamp(6000, 0)  
 { "id" : 29868 } -->> { "id" : 37149 } on : shard0001 Timestamp(7000, 0)  
 { "id" : 37149 } -->> { "id" : 48037 } on : shard0000 Timestamp(8000, 0)  
 { "id" : 48037 } -->> { "id" : { $maxKey : 1 } } on : shard0000 Timestamp(4000, 1)  
 { "_id" : "records", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "fracns", "partitioned" : false, "primary" : "shard0000" }

备注：注意最后一行，库 fracns 的主节点已变为 shard0000 了。

2.6 Finalize the Migration

1
2
3
4
5
6
7
 
mongos> db.runCommand({removeshard:"redhatB.example.com:5283"});  
{  
 "msg" : "removeshard completed successfully",  
 "state" : "completed",  
 "shard" : "shard0002",  
 "ok" : 1  
}

备注：在转移 shard 节点数据之后，再次执行 removeshard 命令，收到以上信息，表示删除 shard 节点成功。

三：验证

3.1 查看 Shard Cluster 状态

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
 
mongos> sh.status();  
--- Sharding Status ---  
 sharding version: { "_id" : 1, "version" : 3 }  
 shards:  
 { "_id" : "shard0000", "host" : "redhatB.example.com:5281" }  
 { "_id" : "shard0001", "host" : "redhatB.example.com:5282" }  
 databases:  
 { "_id" : "admin", "partitioned" : false, "primary" : "config" }  
 { "_id" : "test", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "francs", "partitioned" : true, "primary" : "shard0000" }  
 francs.test_1 chunks:  
 shard0000 4  
 shard0001 3  
 { "id" : { $minKey : 1 } } -->> { "id" : 1 } on : shard0000 Timestamp(5000, 0)  
 { "id" : 1 } -->> { "id" : 7282 } on : shard0001 Timestamp(5000, 1)  
 { "id" : 7282 } -->> { "id" : 15651 } on : shard0001 Timestamp(4000, 3)  
 { "id" : 15651 } -->> { "id" : 29868 } on : shard0000 Timestamp(6000, 0)  
 { "id" : 29868 } -->> { "id" : 37149 } on : shard0001 Timestamp(7000, 0)  
 { "id" : 37149 } -->> { "id" : 48037 } on : shard0000 Timestamp(8000, 0)  
 { "id" : 48037 } -->> { "id" : { $maxKey : 1 } } on : shard0000 Timestamp(4000, 1)  
 { "_id" : "records", "partitioned" : false, "primary" : "shard0000" }  
 { "_id" : "fracns", "partitioned" : false, "primary" : "shard0000" }

备注：上面已经没有 “redhatB.example.com:5283” 节点信息了。

3.2 验证集合 test_1 数据

1
2
3
4
 
mongos> use francs;use francs;  
switched to db francs  
mongos> db.test_1.count();db.test_1.count();  
100000

四 停 “redhatB.example.com:5283” 节点

1
2
3
4
5
 
[shard@redhatB ~]$ ps -ef | grep 5283  
shard 18089 17986 0 14:23 pts/2 00:00:00 grep 5283  
shard 25864 1 0 02:51 ? 00:04:52 mongod -f /pgdata_xc/shard/shard3/shard3_5283.conf  

[shard@redhatB ~]$ kill 25864

备注：以上就是完整的删除节点步骤。

五 附

primary shard ：如果某个库已开启 sharding 功能，并将未分片的集合存储在一个单一的 shard 节点上，那么这个 shard 节点被称为这个库的 primary shard。

六 参考

http://docs.mongodb.org/manual/tutorial/remove-shards-from-cluster/
http://docs.mongodb.org/manual/reference/commands/#removeShard
http://docs.mongodb.org/manual/reference/glossary/#term-primary-shard

原创文章，作者：745907710，如若转载，请注明出处：https://blog.ytso.com/237930.html

赞 (0)
0
上一篇 2022年1月29日
下一篇 2022年1月29日

相关推荐

发表回复

登录后才能评论