ConfigSvr组统一使用默认端口27017（后面发现让Config Server的官方端口是27019）。
修改配置文件(configsvr01、mix01、mix02) [root@configsvr01 ~]# cp /etc/mongod.conf{,.bak} [root@configsvr01 ~]# vi /etc/mongod.conf [root@configsvr01 ~]# tail /etc/mongod.conf # Edit sections below net: port: 27017 bindIp: 127.0.0.1,configsvr01 replication: replSetName: configset sharding: clusterRole: configsvr [root@configsvr01 ~]# [root@mix01 ~]# tail /etc/mongod.conf .... bindIp: 127.0.0.1,mix01 .... [root@mix01 ~]# [root@mix02 ~]# tail /etc/mongod.conf .... bindIp: 127.0.0.1,mix02 .... [root@mix02 ~]# 设定防火墙(configsvr01、mix01、mix02) [root@configsvr01 ~]# firewall-cmd --permanent --add-service=mongodb success [root@configsvr01 ~]# firewall-cmd --reload success [root@configsvr01 ~]# 初始configset组（configsvr01）标准端口的话就不用单独指定了。
[root@configsvr01 ~]# systemctl start mongod.service [root@configsvr01 ~]# mongo > rs.initiate({ ... _id: "configset", ... configsvr: true, ... members: [ ... { _id : 0, host : "configsvr01" }, ... { _id : 1, host : "mix01" }, ... { _id : 2, host : "mix02" }] }) { "ok" : 1, "$gleStats" : { "lastOpTime" : Timestamp(1601776109, 1), "electionId" : ObjectId("000000000000000000000000") }, "lastCommittedOpTime" : Timestamp(0, 0), "$clusterTime" : { "clusterTime" : Timestamp(1601776109, 1), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } }, "operationTime" : Timestamp(1601776109, 1) } configset:SECONDARY> configset:PRIMARY> [root@mix01 ~]# mongo configset:SECONDARY> [root@mix02 ~]# mongo configset:SECONDARY>ShardSet1组统一使用端口27016
初始化实例（shard1st01、mix01、mix02）单机多实例初始化脚本参考：https://blog.csdn.net/weixin_42480750/article/details/108914266
[root@shard1st01 ~]# head -12 setup_another_mongodb_instance.sh #!/bin/bash # Setup Another Mongodb Instance # -- all you need to do is defining two vars below # # suffix : distinguish from standard mongodb instance # e.g standard mongodb instance conf-file : /etc/mongod.conf # this new created instance conf-file : /etc/mongod-new.conf : which will add suffix 'new' # mongodb-port : distinguish from the standard port 27017 # suffix=sh1 mongodb_port=27016 [root@shard1st01 ~]# sh setup_another_mongodb_instance.sh Setup work seems done! Now you can either check the setup log file "/tmp/setup_multiple_mongodb_instance_2020-10-04.log" to see if had something wrong. Or just start the service directly with command : "systemctl start mongod-sh1.service" Then log into mongodb with command : "mongo --port 27016" [root@shard1st01 ~]# 修改配置文件（shard1st01、mix01、mix02） [root@shard1st01 ~]# cp /etc/mongod-sh1.conf{,.bak} [root@shard1st01 ~]# vi /etc/mongod-sh1.conf [root@shard1st01 ~]# cat /etc/mongod-sh1.conf systemLog: destination: file logAppend: true path: /var/log/mongodb-sh1/mongod.log storage: dbPath: /var/lib/mongo-sh1 journal: enabled: true processManagement: fork: true pidFilePath: /var/run/mongodb-sh1/mongod.pid timeZoneInfo: /usr/share/zoneinfo # sections below need to change net: port: 27016 bindIp: 127.0.0.1,shard1st01 sharding: clusterRole: shardsvr replication: replSetName: sh1 [root@shard1st01 ~]# [root@mix01 ~]# cat /etc/mongod-sh1.conf .... bindIp: 127.0.0.1,mix01 .... [root@mix01 ~]# [root@mix02 ~]# cat /etc/mongod-sh1.conf .... bindIp: 127.0.0.1,mix02 .... [root@mix02 ~]# 设定防火墙（shard1st01、mix01、mix02） [root@shard1st01 ~]# firewall-cmd --permanent --add-port=27016/tcp success [root@shard1st01 ~]# firewall-cmd --reload success [root@shard1st01 ~]# 初始化sh1组（shard1st01） [root@shard1st01 ~]# systemctl start mongod-sh1.service [root@shard1st01 ~]# mongo --port 27016 > rs.initiate({ ... _id : "sh1", ... members: [ ... { _id : 0, host : "shard1st01:27016" }, ... { _id : 1, host : "mix01:27016" }, ... { _id : 2, host : "mix02:27016" }] }) { "ok" : 1, "$clusterTime" : { "clusterTime" : Timestamp(1601778520, 1), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } }, "operationTime" : Timestamp(1601778520, 1) } sh1:SECONDARY> sh1:PRIMARY> [root@mix01 ~]# mongo --port 27016 sh1:SECONDARY> [root@mix02 ~]# mongo --port 27016 sh1:SECONDARY>ShardSet1组统一使用端口27015
初始化实例（shard2nd01、mix01、mix02） [root@shard2nd01 ~]# head -12 setup_another_mongodb_instance.sh #!/bin/bash # Setup Another Mongodb Instance # -- all you need to do is defining two vars below # # suffix : distinguish from standart mongodb instance # e.g standard mongodb instance conf-file : /etc/mongod.conf # this new created instance conf-file : /etc/mongod-new.conf : which will add suffix 'new' # mongodb-port : distinguish from the standard port 27017 # suffix=sh2 mongodb_port=27015 [root@shard2nd01 ~]# sh setup_another_mongodb_instance.sh Setup work seems done! Now you can either check the setup log file "/tmp/setup_multiple_mongodb_instance_2020-10-04.log" to see if had something wrong. Or just start the service directly with command : "systemctl start mongod-sh2.service" Then log into mongodb with command : "mongo --port 27015" [root@shard2nd01 ~]# 修改配置文件（shard2nd01、mix01、mix02） [root@shard2nd01 ~]# vi /etc/mongod-sh2.conf [root@shard2nd01 ~]# cat /etc/mongod-sh2.conf systemLog: destination: file logAppend: true path: /var/log/mongodb-sh2/mongod.log storage: dbPath: /var/lib/mongo-sh2 journal: enabled: true processManagement: fork: true pidFilePath: /var/run/mongodb-sh2/mongod.pid timeZoneInfo: /usr/share/zoneinfo # sections below need to change net: port: 27015 bindIp: 127.0.0.1,shard2nd01 sharding: clusterRole: shardsvr replication: replSetName: sh2 [root@shard2nd01 ~]# [root@mix01 ~]# cat /etc/mongod-sh2.conf .... bindIp: 127.0.0.1,mix01 .... [root@mix01 ~]# [root@mix02 ~]# cat /etc/mongod-sh2.conf .... bindIp: 127.0.0.1,mix02 .... [root@mix02 ~]# 设定防火墙（shard2nd01、mix01、mix02） [root@shard2nd01 ~]# firewall-cmd --permanent --add-port=27015/tcp success [root@shard2nd01 ~]# firewall-cmd --reload success [root@shard2nd01 ~]# 初始化sh1组（shard2nd01） [root@shard2nd01 ~]# systemctl start mongod-sh2.service [root@shard2nd01 ~]# mongo --port 27015 > rs.initiate({ ... _id : "sh2", ... members: [ ... { _id : 0, host : "shard2nd01:27015" }, ... { _id : 1, host : "mix01:27015" }, ... { _id : 2, host : "mix02:27015" }] }) { "ok" : 1, "$clusterTime" : { "clusterTime" : Timestamp(1601779334, 1), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } }, "operationTime" : Timestamp(1601779334, 1) } sh2:SECONDARY> sh2:PRIMARY> [root@mix01 ~]# mongo --port 27015 > sh2:SECONDARY> [root@mix02 ~]# mongo --port 27015 > sh2:SECONDARY>MongosRouter统一使用端口27000
初始化实例（mongos01、mix01）这里仍旧使用shell脚本模板来进行实例初始化搭建工作，但有需要额外修改的地方。
[root@mongos01 ~]# head -12 setup_another_mongodb_instance.sh #!/bin/bash # Setup Another Mongodb Instance # -- all you need to do is defining two vars below # # suffix : distinguish from standart mongodb instance # e.g standard mongodb instance conf-file : /etc/mongod.conf # this new created instance conf-file : /etc/mongod-new.conf : which will add suffix 'new' # mongodb-port : distinguish from the standard port 27017 # suffix=router mongodb_port=27000 [root@mongos01 ~]# sh setup_another_mongodb_instance.sh Setup work seems done! Now you can either check the setup log file "/tmp/setup_multiple_mongodb_instance_2020-10-04.log" to see if had something wrong. Or just start the service directly with command : "systemctl start mongod-router.service" Then log into mongodb with command : "mongo --port 27000" [root@mongos01 ~]# 修改Mongos服务启动文件（mongos01、mix01）MongosRouter是通过命令mongos而不是mongod启动的。
[root@mongos01 ~]# sed 's#/usr/bin/mongod#/usr/bin/mongos#' /etc/systemd/system/mongod-router.service [Unit] Description=MongoDB Database Server Documentation=https://docs.mongodb.org/manual After=network.target [Service] User=mongod Group=mongod Environment="OPTIONS=-f /etc/mongod-router.conf" ExecStart=/usr/bin/mongos $OPTIONS ExecStartPre=/usr/bin/mkdir -p /var/run/mongodb-router ExecStartPre=/usr/bin/chown mongod:mongod /var/run/mongodb-router ExecStartPre=/usr/bin/chmod 0755 /var/run/mongodb-router PermissionsStartOnly=true PIDFile=/var/run/mongodb-router/mongod.pid Type=forking LimitFSIZE=infinity LimitCPU=infinity LimitAS=infinity LimitNOFILE=64000 LimitNPROC=64000 LimitMEMLOCK=infinity TasksMax=infinity TasksAccounting=false [Install] WantedBy=multi-user.target [root@mongos01 ~]# sed -i 's#/usr/bin/mongod#/usr/bin/mongos#' /etc/systemd/system/mongod-router.service [root@mongos01 ~]# 修改配置文件（mongos01、mix01） [root@mongos01 ~]# cp /etc/mongod-router.conf{,.bak} [root@mongos01 ~]# vi /etc/mongod-router.conf [root@mongos01 ~]# cat /etc/mongod-router.conf systemLog: destination: file logAppend: true path: /var/log/mongodb-router/mongod.log # storage section removed processManagement: fork: true # fork and run in background pidFilePath: /var/run/mongodb-router/mongod.pid timeZoneInfo: /usr/share/zoneinfo # sections below need to change net: port: 27000 bindIp: 127.0.0.1,mongos01 # default config server port is 27019 sharding: configDB: configset/configsvr01:27017,mix01:27017,mix02:27017 [root@mongos01 ~]# [root@mix01 ~]# cat /etc/mongod-router.conf .... bindIp: 127.0.0.1,mix01 .... [root@mix01 ~]# 开放防火墙（mongos01、mix01） [root@mongos01 ~]# firewall-cmd --permanent --add-port=27000/tcp success [root@mongos01 ~]# firewall-cmd --reload success [root@mongos01 ~]# 启动服务（mongos01、mix01） [root@mongos01 ~]# systemctl start mongod-router.service [root@mongos01 ~]# systemctl status mongod-router.service .... Active: failed (Result: signal) since Sun 2020-10-04 12:09:06 CST; 6s ago .... [root@mongos01 ~]# journalctl -xe Oct 04 12:09:10 mongos01 setroubleshoot[4212]: SELinux is preventing ftdc from create access on the directory mongod.diagnostic.data. For complete SELinux messages run: sealert -l 5c0dff1f-4191-413d-9223-2bc2433> Oct 04 12:09:10 mongos01 platform-python[4212]: SELinux is preventing ftdc from create access on the directory mongod.diagnostic.data. ***** Plugin catchall (100. confidence) suggests ************************** If you believe that ftdc should be allowed create access on the mongod.diagnostic.data directory by default. Then you should report this as a bug. You can generate a local policy module to allow this access. Do allow this access for now by executing: # ausearch -c 'ftdc' --raw | audit2allow -M my-ftdc # semodule -X 300 -i my-ftdc.pp [root@mongos01 ~]# ausearch -c 'ftdc' --raw | audit2allow -M my-ftdc ******************** IMPORTANT *********************** To make this policy package active, execute: semodule -i my-ftdc.pp [root@mongos01 ~]# semodule -i my-ftdc.pp [root@mongos01 ~]# systemctl restart mongod-router.service [root@mongos01 ~]# 添加分片sh1（mongos01）MongosRouter是互通的，在一个上面添加分片即可，与另外一个稍后搭建高可用。
[root@mongos01 ~]# mongo --port 27000 mongos> sh.addShard( "sh1/shard1st01:27016,mix01:27016,mix02:27016") { "ok" : 0, "errmsg" : "Could not find host matching read preference { mode: \"primary\" } for set sh1", "code" : 133, "codeName" : "FailedToSatisfyReadPreference", "operationTime" : Timestamp(1601786011, 1), "$clusterTime" : { "clusterTime" : Timestamp(1601786011, 1), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } } } [root@mongos01 ~]# journalctl -xe Oct 04 14:39:08 mongos01 setroubleshoot[4879]: SELinux is preventing Replica.xecutor from name_connect access on the tcp_socket port 27016. For complete SELinux messages run: sealert -l a0a055aa-f049-445b-9414-7> Oct 04 14:39:08 mongos01 platform-python[4879]: SELinux is preventing Replica.xecutor from name_connect access on the tcp_socket port 27016. ***** Plugin connect_ports (92.2 confidence) suggests ********************* If you want to allow Replica.xecutor to connect to network port 27016 Then you need to modify the port type. Do # semanage port -a -t PORT_TYPE -p tcp 27016 where PORT_TYPE is one of the following: dns_port_t, dnssec_port_t, kerberos_port_t, mongod_port_t, ocsp_port_t. [root@mongos01 ~]# semanage port -a -t mongod_port_t -p tcp 27016 [root@mongos01 ~]# [root@configsrv01 ~]# journalctl -xe Oct 04 14:35:20 configsrv01 setroubleshoot[3099]: SELinux is preventing Replica.xecutor from name_connect access on the tcp_socket port 27016. For complete SELinux messages run: sealert -l 28e0faa5-f5c1-49bf-a7c> Oct 04 14:35:20 configsrv01 platform-python[3099]: SELinux is preventing Replica.xecutor from name_connect access on the tcp_socket port 27016. ***** Plugin connect_ports (92.2 confidence) suggests ********************* If you want to allow Replica.xecutor to connect to network port 27016 Then you need to modify the port type. Do # semanage port -a -t PORT_TYPE -p tcp 27016 where PORT_TYPE is one of the following: dns_port_t, dnssec_port_t, kerberos_port_t, mongod_port_t, ocsp_port_t. [root@configsrv01 ~]# semanage port -a -t mongod_port_t -p tcp 27016 [root@configsrv01 ~]#这里报了好多端口错误，其实默认27017-27019,28017-28019是SELinux开放的，下次用这些好了。
[root@mongos01 ~]# semanage port -l | grep mongod_port_t mongod_port_t tcp 27016, 27000, 27017-27019, 28017-28019 [root@mongos01 ~]# mongos> sh.addShard( "sh1/shard1st01:27016,mix01:27016,mix02:27016") { "shardAdded" : "sh1", "ok" : 1, "operationTime" : Timestamp(1601793390, 5), "$clusterTime" : { "clusterTime" : Timestamp(1601793390, 5), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } } } mongos> 添加分片sh2（mongos01） [root@configsrv01 ~]# semanage port -a -t mongod_port_t -p tcp 27015 [root@configsrv01 ~]# [root@mongos01 ~]# semanage port -a -t mongod_port_t -p tcp 27015 [root@mongos01 ~]# mongo --port 27000 mongos> sh.addShard( "sh2/shard2nd01:27015,mix01:27015,mix02:27015") { "shardAdded" : "sh2", "ok" : 1, "operationTime" : Timestamp(1601794253, 5), "$clusterTime" : { "clusterTime" : Timestamp(1601794253, 5), "signature" : { "hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="), "keyId" : NumberLong(0) } } } mongos> 设置SELinux端口访问复制集添加完成后，setroubleshoot疯狂报错，端口都互相开放一下，下次注意使用标准端口。
[root@shard1st01 ~]# semanage port -l | grep mongod_port_t mongod_port_t tcp 27016, 27017-27019, 28017-28019 [root@shard1st01 ~]# semanage port -a -t mongod_port_t -p tcp 27015 [root@shard1st01 ~]# [root@shard2nd01 ~]# semanage port -l | grep mongod_port_t mongod_port_t tcp 27015, 27017-27019, 28017-28019 [root@shard2nd01 ~]# semanage port -a -t mongod_port_t -p tcp 27016 [root@shard2nd01 ~]# 确认另一台MongosRouter（mix01） [root@mix01 ~]# mongo --port 27000 mongos> sh.status() --- Sharding Status --- shards: { "_id" : "sh1", "host" : "sh1/mix01:27016,mix02:27016,shard1st01:27016", "state" : 1 } { "_id" : "sh2", "host" : "sh2/mix01:27015,mix02:27015,shard2nd01:27015", "state" : 1 } mongos>初始化ReplicaSet的时候没有指定，默认所有成员优先级相同，所以会出现主节点恢复后不能自动为主的情况。
[root@shard1st01 ~]# mongo --port 27016 sh1:SECONDARY> [root@mix01 ~]# mongo --port 27016 sh1:PRIMARY> 修改第一分片shard1st01的优先级，使其可用时即为主： sh1:PRIMARY> cfg = rs.conf() { "_id" : "sh1", "members" : [ { "_id" : 0, "host" : "shard1st01:27016", "priority" : 1, "votes" : 1 }, { "_id" : 1, "host" : "mix01:27016", "priority" : 1, "votes" : 1 }, { "_id" : 2, "host" : "mix02:27016", "priority" : 1, "votes" : 1 } ] } sh1:PRIMARY> cfg.members[0].priority = 10 10 sh1:PRIMARY> rs.reconfig(cfg) { "ok" : 1 } sh1:PRIMARY> sh1:SECONDARY> [root@shard1st01 ~]# mongo --port 27016 sh1:SECONDARY> sh1:PRIMARY> 同样配置其它主节点的优先级 [root@configsrv01 ~]# mongo configset:PRIMARY> cfg = rs.conf() configset:PRIMARY> cfg.members[0].priority = 10 configset:PRIMARY> rs.reconfig(cfg) configset:PRIMARY> [root@mix02 ~]# mongo --port 27015 sh2:PRIMARY> cfg = rs.conf() sh2:PRIMARY> cfg.members[0].priority = 10 sh2:PRIMARY> rs.reconfig(cfg) sh2:PRIMARY> sh2:SECONDARY>