Installing and Configuring Tomcat 6 and Apache Solr for Drupal 6 on Ubuntu

Installation procedure

While there are several ways to install Tomcat 6 and Apache Solr, we will use the repository version to gain the benefit of automatic updates.

What is needed:

1. Tomcat6 as Servlet container

   sudo apt-get install tomcat6
sudo apt-get install tomcat6-admin

2. Apache Solr Search Server

   sudo apt-get install solr-tomcat

Once everything has been correctly installed, you should see the message, “It works!” at http://localhost:8080 and “Welcome to Solr!” at http://localhost:8080/solr/

Configuring Tomcat 6

In the default Tomcat installation, no privileges are created for the Tomcat Manager, so in order to make use of the Tomcat Manager GUI, we still have to create the proper role and a corresponding user.

Add the following lines to your Apache Tomcat 6 configuration file

We’ll edit it with Vim:
sudo vim etc/tomcat6/tomcat-users.xml

Now we need to restart Tomcat:

sudo service tomcat6 restart

Now we can start the manager at http://localhost:8080/manager/html. At this point, you could also deploy Java applets, but for our use case, the manager simply provides an overview of our Solr installations.

Drupal-specific customization

In order to use Solr with Drupal, we still need to make some adjustments to our Solr configuration.
We’ll start by creating a new directory in the file-system:
/opt/solr-tomcat/solr_cocomore_drupal_demo/

We’ll use Vim to create a new file (sudo vim solr_cocomore_drupal_demo.xml) with the following content:

And we need another file (sudo vim solr.xml), containing the following code:

Next we need to create a sub-directory…
sudo mkdir /opt/solr-tomcat/solr_cocomore_drupal_demo/cocomore_drupal_demo

Now we copy the current standard Solr configuration files
cp -R /etc/solr/conf /opt/solr-tomcat/solr_cocomore_drupal_demo/cocomore_drupal_demo/conf

The following files must be present in the conf directory for Apache Solr to function:

• solrconfig.xml
• schema.xml
• elevate.xml
• mapping-ISOLatin1Accent.txt
• protwords.txt
• stopwords.txt
• synonyms.txt

IMPORTANT NOTE!
The files marked in red will be replaced with those included in the ApacheSolr Drupal module.

Now we need to tell Tomcat where our Solr configuration can be found. We will create a symlink for this purpose:

Important: Set correct privileges!

Now, when we visit http://localhost:8080/manager/html, we should see our Solr installation (solr_cocomore_drupal_demo).

Adding and activating the Drupal 6 Apache Solr Modules

The following modules must be installed and activated in our Drupal project:

• Apache Solr Search Integration
• Apache Solr Multilingual

Before we can activate the Apache Solr Search module, we must first install the SolrPhpClient in the ApacheSolr module’s path, ideally with:
svn checkout -r22 http://solr-php-client.googlecode.com/svn/trunk/ SolrPhpClient
Note: If you do not have SVN installed, you can download the SolrPhpClient from a variety of sources. See the readme file included with the ApacheSolr module for further information.

Now we should activate the following modules:

• Search (if not already activated)
• Apache Solr framework
• Apache Solr search
• Apache Solr multilingual

Configuring the Apache Solr module in Drupal 6

admin/settings/apachesolr

Since the default settings should match the module, a connection can already be made to Solr. However, we do not want to use the default Solr index, but our own core, so we have some adjustments to make.

Now we set our Apache Solr core

The core consists of Corename and Instance, entered in the Solr path field:
/solr_cocomore_drupal_demo/cocomore_drupal_demo

Other settings - Apache Solr module (Drupal 6)

“Number of items to index per cron run:” A reasonable value is 50. Higher values can result in problems during Cron runs.

In the “Advanced configuration” fieldset: we should now set Apache Solr Search as the default (“Enabled”).

Search Index (before Cron has been run)

In the “Search index” tab, we can now see how many documents are in the index. Of course we don’t have any yet. We can also see the number of items to be sent for indexing. Content is only indexed during Cron runs. So let’s run Cron!

Search Index (before Cron has been run)

After Cron has run, we should see that there are no items left to send and that documents have been indexed. Please note: There is a delay time after adding new content before it is indexed. In this example, the delay is set to two minutes.

Search fields - “Field biases” fieldset

In the “Search fields” tab, we can specify index “weights” for different fields. Higher weight values for a field give it a greater importance in the search results (not to be confused with “weight” in the context of menus or taxonomies where a higher value makes an item appear lower, i.e. “sink”).

Content bias settings

Under “Type biasing and exclusing”, we can also apply weights to different content types or exclude a content type from the search index, altogether.

Congratulations! Apache Solr is now ready for use!