I'm doing some performance benchmarking of Solr and I started with a single big JSON file containing all the docs that I'm sending via curl. The results are fantastic - I'm achieving an indexing rate of about 44,000 docs/sec using this method (these are really small test docs). In the past I have used CSV and adding overwrite=false to the URL increased performance when doing a fresh reindex when I know all the document ids are unique. I tried this with the JSON upload, and nothing seemed to change. Is this supposed to work with the JSON update handler?
Anyway, Solr is doing spectacular against the competition so far. Keep up the great work! --Dave
