- - - - - - - - - - - - - - - - - - - - - - - - - - - - Name: Maxime Subject: Re: performance issue
Storing billions of documents indexed with dataparksearch is rather unbelievabal. Some search speed improvements has been sheduled for upcoming 4.44 version, but 10 queries per second for multimillion database is an utopia for now. The main idea to distribute index using searchd is intended to make ierarchical chain of searchds running, when a searchd at top layer query several searchd at bottom layer, then marge its results received and return no more than ResultsLimit results to next layer. But the working in such mode is in draft and is not optimized, even more, it never was tested in real environment. - - - - - - - - - - - - - - - - - - - - - - - - - - - - Read the full topic here: http://www.dataparksearch.org/cgi-bin/simpleforum.cgi?fid=02;topic_id=1163004804
