Re: Riak Search - Fast Bulk Insert

And, afaik, a single index.xml file with multiple docs should probably be broken up into one file per doc to make better use of the parallelism already mentioned. Regards, Alexander @siculars http://siculars.posthaven.com Sent from my iRotaryPhone > On Aug 14, 2014, at 10:43, "Eric Redmond"

Re: Riak Search - Fast Bulk Insert

Note that the search-cmd is for search pre 2.0, which does not use solr. If you're planning on using the new Solr based search, you'll need to run Riak 2.0, and write an import script as Dmitri pointed out. Eric Redmond, Engineer @ Basho On Thu, Aug 14, 2014 at 7:38 AM, Dmitri Zagidulin

Re: Riak Search - Fast Bulk Insert

Hi Mark, The best way to bulk load objects into Riak (and into Solr) is to take advantage of Riak's parallelism. Spin up a bunch of worker threads (and have them share a pool of connections) and have them issue parallel concurrent puts to all of the nodes in a cluster (you can either use something

Riak Search - Fast Bulk Insert

Hello What's the fastest way (best practice) to insert 20 million documents into a Riak Search index? search-cmd solr my_bucket /insert.xml For a proof-of-concept I've create a file (index.html) containing 100,000 documents: : Thanks Mark Thomas | Software Engineer | Equifax UK p: +4