Monday 14 May 2012 8:34:25 pm
After completing this tutorial, you should be able to set up a single, multi-core eZ Find instance with eZ Tika. This will allow you to index multiple eZ Publish sites along with virtually any file type, as well as take advantage of eZ Find’s extended searching and sorting functionality.
The multi-core setup variant is useful even if you just have a single eZ Publish site – it doesn’t require any extra resources, and you have the ability to easily add more sites to the search server when the need arises.
eZ Find is a search extension that integrates the content management features of eZ Publish with the powerful, Lucene-based Apache Solr search server. Once installed and configured, it provides faster and more extensive search facilities as well as functionality to overcome some limitations in eZ Publish, such as selecting and sorting on an attribute common to more than one content class.
The eZ Tika extension is a binary file plugin and wrapper for the Apache Tika toolkit, which detects and extracts metadata and structured text content from various document types such as PDF, Word, Excel, ODF, and many more. Used together with eZ Find, virtually all content in an eZ Publish system can be indexed, categorised and searched.
The result is a search server that features enterprise-grade speed and features – the open source way!
It can be tricky to install and configure eZ Find properly, since several technologies are involved and there are some pitfalls to avoid. This tutorial aims to provide a step-by-step guide to the process, as well as give some glimpses into the workings of Solr and how eZ Find integrates it with eZ Publish.
You should be familiar with the Linux command line and package administration, and have a good working knowledge of setting up and using eZ Publish version 4.x (or Community Project releases from 2011 and onward).
Reference platform for this tutorial:
Although this tutorial uses Debian 6 as a reference platform, it should be applicable to other Linux distributions with only minor adaptations. All commands are assumed to be issued as root unless otherwise noted.