I have recently resumed hosting my own blog. My first absence from blogging was the result of giving up on seemingly endless security patches required to keep my Drupal blog alive. When I started using the free as in beer service blogspot.com, I thought I had found the solution to my dearth of blogging. Blogspot was used by several technical people I had seen give presentations, and it was run by Google who I was an embarrassingly huge fan of. Without much hassle, I found it was sufficient to publish any technical notes I accumulated and wanted to share. However, recently I began exploring alternatives for reasons I outline below.

Introduction

I’ve worked with the Apache Jackrabbit implementation of the Java Content Repository (also known as JCR or JSR-170) for some time now, and found it was a bit confusing to get a load balanced implementation. There are plenty of guides and documentation on the Jackrabbit wiki but piecing them together in a way that makes sense took a significant effort. The purpose of this blog post is to describe my approach in the hope that it may make it easier for others with the same goal in mind. I have created a GitHub project with some of the key configuration files included, please feel free to check it out to refer to it as I go along.

I recently replaced my home server with some Apple hardware and instead of having to buy a rather costly Server version of Snow Leopard, Lion had just come out, as well as an App store Server upgrade option. I was delighted to hear that OSX Lion Server shipped with the PostgreSQL database that I use for my application. I know that some people are probably surprised by this, but I think it is an excellent choice over MySQL which I try to avoid.

While setting up a Tomcat that would be handling authentication in the container with an OpenAM instance, I had a hard time finding documentation on where the user’s session attributes would go after they were redirected back to the site. I found this mailing list entry which allowed me to figure the rest out. It was also confusing why it was giving me a 403 when I logged into OpenAM. I will address both of the problems with the solutions I found. For clarity, I am using OpenAM installed from openam_snapshot_952.zip. I say this because I have noticed that the menu items tend to vary wildly between versions.

I run my development environment in Grails almost religiously on HQSLDB. Up until now the only downside was seeing the DB layout before I pushed up to another RDBMS. Now that I’m working with several different plugins, I want to see how they will effect my schema a little earlier. I found this post which seems to be linked from the Grails documentation, but I don’t really want anything fancy (actually, I just don’t want another bloated SQL Tool on my machine). However, I found this post which provided a solution that I like for a simple popup window with a query editor and schema browser. Now in my BootStrap.groovy file I have the following:

1
2
3
4
5
6
7

if("development".equals(GrailsUtil.environment)) {
       // ... various test data
      Thread.start{
          String[] databaseManagerOptions = ['--url', 'jdbc:hsqldb:mem:devDB', '--noexit'];
         org.hsqldb.util.DatabaseManagerSwing.main(databaseManagerOptions)
       }
}

You might notice the command line arguments I passed it, that’s because otherwise HSQLDB’s manager class decides it would be a good idea to call System.exit() when you close the UI. I completely disagree in this case :) Now I can see what the schema looks like and verify that my domain objects are interacting the way I expected them to.