I’m following up on my CFSummit 2013 blog post, a few people have asked me about this, so I wanted to make a page dedicated to this topic.
ColdFusion 10 ships with Solr version 3.4 (Solr Specification Version: 220.127.116.112.02.02.15.09.05 ). It looks like they got the source code from the git repository on February 2nd, 2012. In my application that I was writing, I was indexing a huge amount of data (full text of PDF files). The solr instance bundled within ColdFusion couldn’t handle the amount of data that I was passing it. I don’t remember the exact bug, but I did indeed find that the bug was fixed in one of the newer versions of solr. I emailed Rakshith and asked him if he had documentation for updating Solr and he told me that unfortunately there wasn’t a way to do it for reasons that I quoted in the CF Summit blog post.
Solr is now on version 4.5.1. From Version 3 to version 4 they made significant improvements, listed in the official docs here: http://docs.lucidworks.com/display/solr/Major+Changes+from+Solr+3+to+Solr+4
I decided try to install solr myself. It sounded like a scary task, but it was extremely easy to leverage the power of solr with an open source library, cfsolrlib. I found it almost as easy to use as the native <cfsearch> and <cfindex> tags. The library was written by Shannon Hicks and is on riaforge: http://cfsolrlib2.riaforge.org. It’s repository is located on github at https://github.com/iotashan/cfsolrlib.
I found that using “real solr” helped me understand how solr actually works. Prior to installing solr from scratch I was telling my clients that “solr is a black box!” but by installing the full version of solr I was able to have control over things such as word stemming, word matching, etc.
Don’t be afraid to use the “real” version of solr!