MongoDB 副本集分片集群一分片config库主机断电导致该分片config库无法启动

    2018年，1月22日，由于机房意外断电，一分片config服务器断电导致该config库无法启动，相关报错如下：



关键报错信息：
2018-01-22T15:18:43.863+0800 I -[initandlisten] Detected data file in D:\mongodata\cfg created
by the 'wiredTiger' storage
 engine,so setting the active storage engine to 'wiredTiger'.
2018-01-22T15:18:43.865+0800 W- [initandlisten] Detected
unclean shutdown -D:\mongodata\cfg\mongod.lock is not empty.
2018-01-22T15:18:43.866+0800 w STORAGE [initandlisten] Recovering data from the last clean
checkpoint.
根据网上查到的处理方法：
1、关闭启动窗口
2、删除D:\mongodata\cfg\mongod.lock，
3、执行修复操作：D:\MongoDB\Server\3.2\bin\mongod
--configsvr --replSet configReplSet --port 27001 
--dbpath D:\mongodata\cfg  --repair
2018-01-22T15:46:50.530+0800 I STORAGE  [initandlisten] Repairing collection loc
al.system.replset
2018-01-22T15:46:50.531+0800 I STORAGE  [initandlisten] Verify succeeded on uri
table:collection-4--3441646246172970043. Not salvaging.
2018-01-22T15:46:50.545+0800 I INDEX    [initandlisten] build index on: local.sy
stem.replset properties: { v: 1, key: { _id: 1 }, name: "_id_", ns: "local.syste
m.replset" }
2018-01-22T15:46:50.545+0800 I INDEX    [initandlisten]          building index
using bulk method
2018-01-22T15:46:50.569+0800 I STORAGE  [initandlisten] finished checking dbs
2018-01-22T15:46:50.569+0800 I CONTROL  [initandlisten] now exiting
2018-01-22T15:46:50.569+0800 I NETWORK  [initandlisten] shutdown: going to close
 listening sockets...
2018-01-22T15:46:50.570+0800 I NETWORK  [initandlisten] shutdown: going to flush
 diaglog...
2018-01-22T15:46:50.570+0800 I NETWORK  [initandlisten] shutdown: going to close
 sockets...
2018-01-22T15:46:50.571+0800 I STORAGE  [initandlisten] WiredTigerKVEngine shutt
ing down
2018-01-22T15:46:50.769+0800 I STORAGE  [initandlisten] shutdown: removing fs lo
ck...
2018-01-22T15:46:50.770+0800 I CONTROL  [initandlisten] dbexit:  rc: 0

4、修复操作没有明显的报错，重新启动config库，问题依然存在，报错与上述提到的日志一致

由于是副本集分片集群，且该config库只是一个片的config库，该副本集分片集群一共有3台config库，于是决定删除该config库
的data目录D:\mongodata\cfg下的所有数据文件以从集群其他节点重新同步，该config库重新同步数据后恢复正常。