On Mar 9, 2013, at 11:54 AM, Paul Jungwirth wrote:
Hello,
I'm running a specialized search engine that indexes a few tens of millions
of web pages, keeping everything in Postgres, and one problem I'm starting to
see is poor cache hit rates. My database has two or three tables just for the
Well, what problem exactly are you trying to solve?
Having large tables itself isn't a problem, but it often
tends to imply other things that might be problematic:
I'm trying to troubleshoot a very low cache hit rate as returned by this query:
SELECT sum(heap_blks_read) as heap_read,
Hello,
I'm running a specialized search engine that indexes a few tens of millions
of web pages, keeping everything in Postgres, and one problem I'm starting
to see is poor cache hit rates. My database has two or three tables just
for the text of the scraped pages, with one row every time a page