replSet error RS102 too stale to catch up
辅助副本成员需要关机增加内存，很早就提前关机，等第二天服务器起来后，去查看rs.status()信息，会捕获到"stateStr" : "RECOVERING"。其实早已超过了维护窗口时间。这时，如果运气好的话，会看到error RS102 too stale to catch up错误。而该错误是每十分钟会捕获到一次。从mongodb的日志文件可以看到：
Thu Mar 26 20:50:48.561 [rsBackgroundSync] replSet error RS102 too stale to catch up, at least from dal05mgo13.sl.dx:27017
Thu Mar 26 20:50:48.561 [rsBackgroundSync] replSet error RS102 too stale to catch up
Thu Mar 26 21:00:49.632 [rsBackgroundSync] replSet error RS102 too stale to catch up, at least from dal05mgo13.sl.dx:27017
Thu Mar 26 21:00:49.632 [rsBackgroundSync] replSet error RS102 too stale to catch up
Thu Mar 26 21:10:50.701 [rsBackgroundSync] replSet error RS102 too stale to catch up, at least from dal05mgo12.sl.dx:27017
Thu Mar 26 21:10:50.701 [rsBackgroundSync] replSet error RS102 too stale to catch up
Thu Mar 26 21:20:51.770 [rsBackgroundSync] replSet error RS102 too stale to catch up, at least from dal05mgo13.sl.dx:27017
Thu Mar 26 21:20:51.770 [rsBackgroundSync] replSet error RS102 too stale to catch up
确认维护窗口大小
查看主副本的维护窗口和oplog大小：
rs_main:PRIMARY> db.printReplicationInfo()
configured oplog size: 23552MB
log length start to end: 38133secs (10.59hrs)
oplog first event time: Mon Mar 30 2015 12:00:13 GMT-0500 (CDT)
oplog last event time: Mon Mar 30 2015 22:35:46 GMT-0500 (CDT)
now: Mon Mar 30 2015 22:35:46 GMT-0500 (CDT)
可以看到维护窗口为10个小时，oplog大小为23G。
辅助副本的维护窗口大小最好跟主副本成员保持一致，让辅助副本能保存从主副本同步过来的oplog。
如何重建oplog？
以standalone方式启动，执行如下删除和创建脚本：
> use local
> db.oplog.rs.drop()
> db.createCollection("oplog.rs", {"capped" : true, "size" : 23 * 1024 * 1024 * 1024})
或
> db.runCommand( { create: "oplog.rs", capped: true, size: (23 * 1024 * 1024 * 1024) } )
replSet error RS102 too stale to catch up如何处理？
副本集错误RS102太陈旧而无法跟上。说明此节点由于脱离副本集时间太长，已经跟不上整个集群了。需要重新同步该辅助副本成员。可参考：
http://docs.mongodb.org/manual/tutorial/resync-replica-set-member/
官方手册建议，通过移除数据实施初始化同步（initial sync）来完全重新同步成员。
MongoDB提供了两种实施初始化同步的方法：