Semantic Search - Faceted Search and Web 3.0 for Drupal

Starting a new project over at Drupal.org for running Drupal on top of Vertica.

Why? I like MySQL, and have used it for years - and it's been reliable, fast, and useful for just about every application I've needed to use - except one: as an RDF backend.

This project isn't RDF related however, and Vertica has some promising features that could make Drupal's heavy dependence on database calls (rather inefficient at times) more performant and scalable. VerticaDB offers some out-of-the box performance improvements that would theoretically improve Drupal performance dramatically. As of this writing, I may don't know yet if I'll use the Vertica JDBC driver and run Drupal inside Java using Quercus, or use an ODBC connection.

But that's beyond the scope of this tutorial.

This article covers only the installation and testing of Vertica in a development environment accessible to those without a bunch of spare Linux servers. Vertica officially supports Suse and Red Hat Enterprise Linux, but Vertica will run on CentOS and Fedora.

Parallels doesn't like 64 bit Linux flavors (you may have more luck with VMWare), and CentOS5 had a kernel panic. I like Parallels, and I like VMWare - they sometimes fill different needs. CentOS 4 installed fine, and the version of VMWare I installed was
vertica_2.0.11-0.i386.RHEL4.rpm . I also installed the Python extension, though I won't be using it for this project.

Steps:

1. installed the rpm
2. configured the hostname to be host01 (Vertica is designed to use many hosts very easily). Note that you can not use localhost
3. set up passwordless SSH access (right, these are meant to be behind a firewall - talking, and let them only talk this way to eachother...)
4. ran the install_vertica tool. This is pretty big script - does a lot of checks, including hostname resolution
5. logged in as dbadmin
6. ran /opt/vertica/bin/adminTools as that user
7. status: acquiring license key

Very straightforward install.