I’m working on a Rails app using aaf and a DRb server to query a Ferret
index. Recent upgrades in the code are using Ferret significantly more,
and we were torquing the DRb server pretty hard until we started caching
some more query-intense fragments. It’s a multi-core server, but the
single DRb server only seems to want to use one thread at a time, not
unlike Mongrel.
As site load increases, the breathing room we won with caching is likely
to shrink. Is the approach described here:
http://www.ruby-forum.com/topic/131816#598404
(and by reference:
http://kalvir.blogspot.com/2007/12/multiple-searchers-on-ferret.html
)
…still the most effective approach for scaling up Ferret/DRb?
Thanks,
pjm