- - - - - - - - - - - - - - - - - - - - - - - - - - - - Name: Jon Subject: Recommendation On Settings
Thanks for a great search system. The cache mode searches millions of documents quickly. However, I have noticed a terrible slowdown in indexing when we start to get this many documents. For a site with multiple millions of documents, what is a recommended setting for WrdFiles, CacheLogWords, CacheLogDels, URLDataFiles, OptimizeAtUpdate (i assume 'no'), OptimizeInterval and OptimizeRatio? Also, would it be quicker to not enable zlib to save CPU time vs file system space. On that note, is there a recommended file system to use? Currently it is using ext3 on LVM. I am looking at use XFS on LVM and just have not done the switch for I am waiting to do some more research on XFS. Maybe reiserfs? Thanks in advance. Also, are there any other conifguration details that would be good for scaling dpsearch to millions of documents? On another note, I still get strange 'access denied' messages that are blamed on mySQL. If I reset the index, the access denied ends up on random keywords. Such, if I search for 'asdf' it give an access denied message and if I search for 'asdf a' 'asdf b' 'asdf quick' .. etc. etc. it works. The key term(s) that get denied change every time I reset the indexed data. - - - - - - - - - - - - - - - - - - - - - - - - - - - - Read the full topic here: http://dataparksearch.org/cgi-bin/simpleforum.cgi?fid=02;post=
