SIFTER Database Creation

The database server doesn't need to be located on the same machine as the web server, though it usually makes sense for them both to be on the same network segment, connected by a fast switch. The minimum requirement is that the web server needs to be able to connect to the database server via Perl DBI. This involves getting the underlying database drivers to work on your system, then installing the appropriate DBD driver for your database to allow access to the data via Perl.

Your next step is to create the database and tables that SIFTER will use. You should already have a database system like Sybase installed. Creating the database will usually be done by your Database Administrator (DBA). The following database systems are supported:

The exact details to create your SIFTER database will vary slightly with each installation. We have provided some scripts and the necessary SQL, but your DBA may have his own way to do the following steps.

Create project database user

For each SIFTER project, you should define a database user which will be used to query and modify SIFTER data. SIFTER uses one database user for all activities - loading data, administering the projects, as well as making queries.

Create project database

For each SIFTER project, you should create a database to contain the SIFTER tables. Space is always a consideration at this point. The vast majority of data in the SIFTER database will be data extracted from the results and not SIFTER overhead. In our experience, an "average" for 1000 analysis results usually requires about ??999??MB of disk space. The database should be large enough to hold at least your initial data. If you are doing a genome scan, 1GB is probably a comfortable size.

Create the tables

In the SIFTER distribution directory, all database related files can be found in the db subdirectory. Under 'db', you'll find a subdirectory for each database engine supported by SIFTER (e.g. 'db/sybase' etc.). In these subdirectories, you will find the SQL needed to create the SIFTER tables. You may wish to modify these slightly to meet your local needs, but these should be very close to meeting your needs.

A Perl script, db/rebuild_sybase_db.plx, is provided to help you run the SQL to create the SIFTER tables. This script is invoked by installdemo.pl and createproject.pl. Be sure to look at this and make sure it will work for your environment. The command to create the SIFTER tables is:

  perl -w db/rebuild_sybase_db.plx YOUR_REALM_NAME

Continue with Demo Installation

Version=$Id: createdb.html,v 1.3 2002/09/13 16:53:28 tpg Exp $