MongoDB副本集--Secondary节点实例恢复

场景描述

MongoDB副本集中有一台Secondary节点出现RECOVERING的状态

状态如下:

    arps:RECOVERING> rs.status()
    {
            "set" : "arps",
            "date" : ISODate("2017-12-22T02:31:58.803Z"),
            "myState" : 3,
            "members" : [
                    {
                            "_id" : 0,
                            "name" : "172.17.4.37:27017",
                            "health" : 1,
                            "state" : 2,
                            "stateStr" : "SECONDARY",
                            "uptime" : 7579839,
                            "optime" : Timestamp(1513909913, 3),
                            "optimeDate" : ISODate("2017-12-22T02:31:53Z"),
                            "lastHeartbeat" : ISODate("2017-12-22T02:31:58.019Z"),
                            "lastHeartbeatRecv" : ISODate("2017-12-22T02:31:57.750Z"),
                            "pingMs" : 0,
                            "syncingTo" : "172.17.4.38:27017",
                            "configVersion" : 1
                    },
                    {
                            "_id" : 1,
                            "name" : "172.17.4.38:27017",
                            "health" : 1,
                            "state" : 1,
                            "stateStr" : "PRIMARY",
                            "uptime" : 7579913,
                            "optime" : Timestamp(1513909913, 3),
                            "optimeDate" : ISODate("2017-12-22T02:31:53Z"),
                            "lastHeartbeat" : ISODate("2017-12-22T02:31:58.051Z"),
                            "lastHeartbeatRecv" : ISODate("2017-12-22T02:31:58.018Z"),
                            "pingMs" : 0,
                            "electionTime" : Timestamp(1506330005, 1),
                            "electionDate" : ISODate("2017-09-25T09:00:05Z"),
                            "configVersion" : 1
                    },
                    {
                            "_id" : 2,
                            "name" : "172.17.4.39:27017",
                            "health" : 1,
                            "state" : 3,
                            "stateStr" : "RECOVERING",//RECOVERING状态，第三个结点出现问题。
                            "uptime" : 7580364,
                            "optime" : Timestamp(1473614444, 2),
                            "optimeDate" : ISODate("2016-09-11T17:20:44Z"),
                            "configVersion" : 1,
                            "self" : true
                    }
            ],
            "ok" : 1
    }

恢复思路:
1.关闭MongoDB故障节点的数据库服务，移除数据目录，启动MongoDB服务，开启自动同步机制，恢复secondary节点。
2.找到另外一个secondary数据节点的快照，关闭写操作。在数据不变化的情况下，获得一致性的备份快照，拷贝至故障节点中，启动MongoDB服务，应用oplog日志。恢复secondary节点。

由于环境数据量小，使用第一种方案。

1.mongodb数据库服务关闭

arps:RECOVERING> use admin
switched to db admin
arps:RECOVERING> db.shutdownServer()

2.删除或者移走数据目录

[root@mongodb data]# mv /opt/data/mongodb /opt/data/mongodb20171222
[root@mongodb data]# mkdir /opt/data/mongodb
[root@mongodb data]# mkdir /opt/data/mongodb/log

3.启动数据库服务且查看状态

[root@mongodb data]#/opt/software/mongodb-linux-x86_64-3.0.1/bin/mongod -f /opt/software/mongodb-linux-x86_64-3.0.1/bin/mongodb.conf

 arps:STARTUP2> rs.status()
{
        "set" : "arps",
        "date" : ISODate("2017-12-22T02:46:52.288Z"),
        "myState" : 5,
        "syncingTo" : "172.17.4.38:27017",
        "members" : [
                {
                        "_id" : 0,
                        "name" : "172.17.4.37:27017",
                        "health" : 1,
                        "state" : 2,
                        "stateStr" : "SECONDARY",
                        "uptime" : 25,
                        "optime" : Timestamp(1513910813, 3),
                        "optimeDate" : ISODate("2017-12-22T02:46:53Z"),
                        "lastHeartbeat" : ISODate("2017-12-22T02:46:51.122Z"),
                        "lastHeartbeatRecv" : ISODate("2017-12-22T02:46:51.114Z"),
                        "pingMs" : 0,
                        "syncingTo" : "172.17.4.38:27017",
                        "configVersion" : 1
                },
                {
                        "_id" : 1,
                        "name" : "172.17.4.38:27017",
                        "health" : 1,
                        "state" : 1,
                        "stateStr" : "PRIMARY",
                        "uptime" : 25,
                        "optime" : Timestamp(1513910813, 3),
                        "optimeDate" : ISODate("2017-12-22T02:46:53Z"),
                        "lastHeartbeat" : ISODate("2017-12-22T02:46:51.127Z"),
                        "lastHeartbeatRecv" : ISODate("2017-12-22T02:46:51.303Z"),
                        "pingMs" : 0,
                        "electionTime" : Timestamp(1506330005, 1),
                        "electionDate" : ISODate("2017-09-25T09:00:05Z"),
                        "configVersion" : 1
                },
                {
                        "_id" : 2,
                        "name" : "172.17.4.39:27017",
                        "health" : 1,
                        "state" : 5,
                        "stateStr" : "STARTUP2",//STARTUP2的状态为:新加入的节点做数据初始化
                        "uptime" : 27,
                        "optime" : Timestamp(0, 0),
                        "optimeDate" : ISODate("1970-01-01T00:00:00Z"),
                        "syncingTo" : "172.17.4.38:27017",
                        "configVersion" : 1,
                        "self" : true
                }
        ],
        "ok" : 1
} 

关于副本集的状态，文献参考如下:https://docs.mongodb.com/v3.0/reference/replica-states/index.html

过了半个小时之后，数据恢复完成，状态日志如下:

    .....................
    2017-12-22T11:27:02.474+0800 I INDEX [rsSync] building index using bulk method
    2017-12-22T11:27:02.475+0800 I INDEX [rsSync] build index done. scanned 75 total records. 0 secs
    2017-12-22T11:27:02.477+0800 I REPL [rsSync] initial sync data copy, starting syncup
    2017-12-22T11:27:02.798+0800 I REPL [rsSync] oplog sync 1 of 3
    2017-12-22T11:27:03.145+0800 I REPL [ReplicationExecutor] syncing from: 172.17.4.38:27017
    2017-12-22T11:27:03.288+0800 I REPL [rsSync] oplog sync 2 of 3
    2017-12-22T11:27:03.289+0800 I REPL [rsSync] initial sync building indexes
    2017-12-22T11:27:03.289+0800 I REPL [rsSync] initial sync cloning indexes for : demo
    2017-12-22T11:27:03.300+0800 I REPL [SyncSourceFeedback] replset setting syncSourceFeedback to 172.17.4.38:27017
    2017-12-22T11:27:03.390+0800 I STORAGE [rsSync] copying indexes for: { name: "ACT_AUTH_LOG", options: {} }
    2017-12-22T11:27:03.391+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_DATA_LOG", options: {} }
    2017-12-22T11:27:03.392+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_ERROR_LOG", options: {} }
    2017-12-22T11:27:03.392+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_EXTERNAL_PACKET", options: {} }
    2017-12-22T11:27:03.393+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_EXTERNAL_PACKET_LOG", options: {} }
    2017-12-22T11:27:03.393+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_JPUSH_LOG", options: {} }
    2017-12-22T11:27:03.394+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_MESSAGE_LOG", options: {} }
    2017-12-22T11:27:03.395+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_REQUEST_LOG", options: {} }
    2017-12-22T11:27:03.395+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_RETRY_MESSAGE", options: {} }
    2017-12-22T11:27:03.395+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_RUN_LOG", options: { capped: true, size: 536870912 } }
    2017-12-22T11:27:03.396+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_SMSEMAIL_LOG", options: {} }
    2017-12-22T11:27:03.396+0800 I STORAGE [rsSync] copying indexes for: { name: "SYSTEM_TIMEOUT_LOG", options: {} }
    2017-12-22T11:27:03.397+0800 I REPL [rsSync] oplog sync 3 of 3
    2017-12-22T11:27:03.406+0800 I REPL [rsSync] initial sync finishing up
    2017-12-22T11:27:03.406+0800 I REPL [rsSync] replSet set minValid=5a3c7b93:3
    2017-12-22T11:27:03.429+0800 I REPL [rsSync] initial sync done
    2017-12-22T11:27:03.474+0800 I REPL [ReplicationExecutor] transition to RECOVERING
    2017-12-22T11:27:03.476+0800 I REPL [ReplicationExecutor] transition to SECONDARY
    .................

节点恢复的状态，如下:

    arps:SECONDARY> rs.status()
    ...............
                    {
                            "_id" : 2,
                            "name" : "172.17.4.39:27017",
                            "health" : 1,
                            "state" : 2,
                            "stateStr" : "SECONDARY",//恢复完成
                            "uptime" : 2500,
                            "optime" : Timestamp(1513913295, 3),
                            "optimeDate" : ISODate("2017-12-22T03:28:15Z"),
                            "syncingTo" : "172.17.4.38:27017",
                            "configVersion" : 1,
                            "self" : true
                    }
    .................