MongoDB游标导致日志文件异常

Question

I am investigating a problem where our MongoDB performace, degrades hideously when I restart the database computer, I'm guessing because the memory mapped files are gone. One thing I found extremely puzzling in my log file was the following (versions of which are repeated many times)

2015-01-14T13:18:59.404-0600 [conn565] query Analitiqa.ScanInfo query: { query: { Node: -6277503169757883360, StartTime: { $gte: new Date(1418882340000) }, EndTime: { $lte: new Date(1421301660000) } }, $hint: { Node: 1, StartTime: 1, EndTime: -1 } } planSummary: IXSCAN { Node: 1, StartTime: 1, EndTime: -1 } cursorid:167701167819 ntoreturn:0 ntoskip:0 nscanned:102 nscannedObjects:101 keyUpdates:0 numYields:97 locks(micros) r:152012 nreturned:101 reslen:243838 3997ms

2015-01-14T13:20:04.819-0600 [conn565] getmore Analitiqa.ScanInfo cursorid:167701167819 ntoreturn:0 keyUpdates:0 numYields:1500 locks(micros) r:4469422 nreturned:1690 reslen:4195825 65415ms

2015-01-14T13:20:23.383-0600 [conn565] getmore Analitiqa.ScanInfo cursorid:167701167819 ntoreturn:0 keyUpdates:0 numYields:390 locks(micros) r:366679 nreturned:456 reslen:1151794 18489ms 2015-01-14T13:20:23.618-0600 [conn565] end connection 127.0.0.1:51230 (15 connections now open)

My reading of the first line is that the query used the correct index, and is returning 101 results (after an unconscionable delay). The next two lines floor me. They appear to indicate that the cursor gotten from the query returns first 1690, and then 456 items, both after enormous delays.

How is that possible if the query was (correctly) returning only 101 results?