Marco, There are many factors that make this a difficult question to answer. How many terms exist in those documents, how many fields, etc. You'll only likely find out the exact parameters for a single Solr instance by actually trying it with your own data.
Having said that, you can break down your index into multiple shards, and then aggregate the results of searches across those shards. I've recently seen a setup that has 10 shards with 4m documents in each shard, for example - total of 40m documents, but aiming for a total of 60m. At that scale there are many issues to watch out for, (memory requirements, impact of index optimisations, etc), but it certainly can be done. Upayavira On Thu, 14 Oct 2010 14:01 +0200, "Marco Ciaramella" <ciaramellama...@gmail.com> wrote: > Hi all, > I am working on a performance specification document on a > Solr/Lucene-based > application; this document is intended for the final customer. My > question > is: what is the maximum number of document I can index assuming 10 or > 20kbytes for each document? > > I could not find a precise answer to this question, and I tend to > consider > that Solr index can be virtually limited only by the JVM, the Operating > System (limits to large file support), or by hardware constraints (mainly > RAM, etc. ... ). > > Thanks > Marco >
