Mustru

  * Front page
  * Overview
     Requirements
  * Install
  * Screenshots
  * Documentation
     FAQ
     User guide
     Related links
     API
  * License
  * Download
  * Evaluation

 
 

Mustru: Install

Linux

  1. Unzip the distribution to the installation directory.
  2. Set the MUSTRU_HOME environment variable to the installation directory
  3. Set the CATALINA_OPTS environment variable to
    export CATALINA_OPTS="$CATALINA_OPTS -DMUSTRU_HOME=$MUSTRU_HOME"
  4. Set the JAVA_HOME to the JDK directory or JRE_HOME to the JRE directory (optional if one of the vars exists)
  5. Add MUSTRU_HOME/bin to your PATH environment variable
  6. Run runOnlineIndex.sh from the bin directory and run Load Tables from the Test menu.
  7. Copy the mustru.war file in the lib directory to your Tomcat webapps directory.

Windows

  1. Run the setup.exe file to copy files to the specified installation directory and set environment vars.
  2. Set the JAVA_HOME to the JDK directory or JRE_HOME to the JRE directory (optional if one of the vars exists)
  3. Run runOnlineIndex.sh from the Start menu under Mustru Services to create an index and run Load Tables from the Test menu.
  4. Copy the mustru.war file in the lib directory to your Tomcat webapps directory.
  5. Tomat configuration: You can use the GUI shown below to configure Mustru and increase heap allocation or set the environment variable MUSTRU_HOME in the startup.bat file and JAVA_OPTS in catalina.bat.


Using Mustru

  1. Build an index - run runOnlineIndex.sh to create / modify an index, options include
    1. A fresh crawl (will create a new index) or an incremental crawl
    2. Duplicates (files the same content) can be skipped
    3. Hidden files (files starting with a '.') can be skipped
    4. Number of threads to create an index (use more than 1 on a multi-processor machine)
    5. A list of directories that will be scanned for indexable files
    6. A list of directories that should not be included in the index
    7. Extract entities (needed for Q&A function alone)
    8. Location of the directory that will contain the Lucene index
    9. Location of the directory that will contain the Berkeley DB
    10. Optional root directory of the local Web server (needed to generate URLs with the localhost in a list of hits)
    Symbolic links are not followed during a directory scan.
  2. Search the index from a browser - start the following URL http://localhost:8080/mustru
You can use Mustru to run queries alone by not extracting entities during the crawl. Memory requirements will be less and indexing will take less time. Extracting entities is memory and computationally intensive. Indexing is much slower, but the time to process a question is reasonable.

Troubleshooting

  1. Tomcat startup error: Tomcat may not start for several reasons. Check the RUNNING.txt file in the main directory for errors. Also check if your firewall is blocking connections.
  2. OutofMemory error: Tomcat starts with an initial heap size which may not be sufficient for Mustru. To increase the heap size, add the following line in the beginning of your bin/catatlina.(sh/bat) file
    set JAVA_OPTS="-Xmx768M -Xms64M" or use the environment variable set CATALINA_OPTS="%CATALINA_OPTS% -Xmx768M -Xms64M"
  3. Could not create DB environment (JE 3.1.0) The Environment directory ... : The Berkeley database directory maybe read only or may not exist. Check for permissions to write to the database.

Copyright © 2007 Mustru Search Services. All rights reserved.