I received 'reduce_overflow_error' when querying views, after entering largish (2-4kb) docs to couch. The problem was due to a specific view, and when I set 'reduce_limit = false' it went away.
Specifically, each of my documents is related to a logical entity in my app, and the purpose of this view is to aggregate information per entity - it maps the documents by the entities to which they are related, and reduces the results (with group=true) so that each entity ends up with its own document. The reduce mainly does counts and sums, but it may also accept quite a bit of text (~4kb) from one of the documents it reduces. I make sure that these blobs of text don't pile up during reduce and rereduce - not more than one will be kept at each stage. Is 4kb too big to keep when reducing? or maybe the problem is that the product of the reduce can be equal to or larger than any of the docs it reduces? Thanks, a.
