Secondary Indexes - Feedback?
mjr at ranney.com
Mon Nov 21 21:44:33 EST 2011
The 2i features are certainly more convenient than building secondary
indexes manually, which is the only alternative until now. I'm a little
concerned about the performance of the coverage queries as our cluster
grows. Building a secondary index manually does scale very nicely as new
nodes are added. But with the new 2i feature, it seems like adding more
computers effectively makes each 2i query do more work.
Also, my initial testing with 2i on a 3-node cluster, n=3 seems like 2i
reads are a lot slower than k/v reads, like maybe 3X slower? I know magic
like this is never free, but that seems like a pretty high price to pay.
On Wed, Nov 16, 2011 at 7:57 AM, Rusty Klophaus <rusty at basho.com> wrote:
> Does the feature help solve your problems? If not, why not? Any concerns?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the riak-users