I dropped the oplog.rs collection and re-created it with a 100GB file.

After 8 hours, I printReplicationInfo shows the wrong data:

aclimasensordb:PRIMARY> db.printReplicationInfo()
configured oplog size:   102400MB
log length start to end: 28.997999999905005secs (0.01hrs)
oplog first event time:  Fri Jan 16 1970 14:24:05 GMT-0500 (EST)
oplog last event time:   Fri Jan 16 1970 14:24:34 GMT-0500 (EST)
now:                     Sat Apr 13 2013 13:30:50 GMT-0400 (EDT)

If I actually query the table, I see that the oplogs are properly there:

aclimasensordb:PRIMARY> db.oplog.rs.find({},

).sort({$natural:-1}).limit(1)
{ "ts" :

}

aclimasensordb:PRIMARY> db.oplog.rs.find({},

).sort({$natural:1}).limit(1)
{ "ts" :

}

t = 1365845252 is 4/13/2013 2:27:32 AM

t = 1365874292 is 4/13/2013 10:31:32 AM