Was doing some benchmarks with search and while thinking how to implement something more-or-less relevant i found this: http://zooie.wordpress.com/2009/07/06/a-comparison-of-open-source-search-engines-and-indexing-twitter/. So I forked and added riak https://github.com/Techmind/opensearch (though building java client took too much time, riak-java-client doesn't have support for search querying and sorlj depends on old libraries).
It is single node test (as other tests use one node too) and i think i couldn't screw up in too many places because test is quite simple. I have to filter-out words like `with|of|and|the|a` because riak complained about `too_many_results`. So results was like: 30-40 min indexing, 20 seconds querying, indexing slower than sqlite and a bit faster querying than sqlite, but much slower than any other search-engine =) I think basho should really discourage people to use riak-search or optimize it somehow at least for something like batch-indexing. -- email: bogu...@gmail.com skype: i.bogunov phone: +7 968 842 5783 Regards, Bogunov Ilya
_______________________________________________ riak-users mailing list riak-users@lists.basho.com http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
