Details
-
Bug
-
Status: Closed
-
Blocker
-
Resolution: Fixed
-
0.2.4, 0.2.3
-
None
Description
During MapReduce indexing, if the number of records in a row exceeds the memory limit and the indexer has to overall to disk the first document added to that tmp index is the prime doc. However if that overflow index goes through several merges the primedoc may not be the first document in the tmp index. If this occurs it causes problems in the row queries once the indexes are brought online in a blur table.