Platform Independence for Building and Installing the RIF

These are some notes I made in March 2018, on getting the RIF installed and running on my Mac. The changes I had to make should at least act as a guide to installing it on Linux or other Unixes in the future.

Note that this is still very much a work in progress.

Tomcat and Postgres

Installing these is straightforward. You can just download and run the relevant installers. I used Homebrew, since I have that installed:

brew install tomcat
brew install postgresql
brew install postgis

For Linux there will be similar commands using apt-get or yum, depending on your distro.

Building the WAR Files

Rather than trying to emulate the Windows java_build.bat script, I just did this manually, with the relevant Maven commands:

mvn clean install

in the root gives us rifServices.war, taxonomy.war, and RIF40.war in the target directories of the corresponding projects: rifServices, taxonomyServices, and rifWebApplication.

I installed the three WAR files using the Tomcat Manager application. I had to increase the maximum allowed size of WAR file in Tomcat, as rifServices.war is bigger than the default maximum, and so wouldn’t deploy. This means editing web.xml in the manager app. I doubled the maximum from 50 to 100MB.

Configuring the Database

The config file postgresql.conf was in /usr/local/var/postgres/. I ran:

initdb /usr/local/var/postgres/ -E utf8

as recommended for example, here. That reported an error, saying the directory already existed and was not empty. I’m not quite sure with the Homebrew run had created it or if I had previously run the above command and forgotten, but the database cluster was installed in that directory.

The next step was to create a script

Database Installation Script

I modelled the Bash script on the existing Windows batch file, rif40_database_install.bat. That script is at rifDatabase/Postgres/production/rif40_database_install.bat. The new one is rifDatabase/Postgres/psql_scripts/rif40_database_install.sh. I have not put it in the production subdirectory because it is not yet production quality.

The major limitations at the moment are that it has several hardcoded values and it doesn’t prompt the user for anything (except a single go/no-go).

But it sets the necessary variables and runs db_create.sql, which works. So at least it is a viable proof of concept.

Tweaks to Get the Database to Create

Note first that the version of db_create.sql is not the one in master on GitHub. Instead it is the “August 2017” one that Peter gave me to run on my PC (rifDemo/August 2017/Postgres/production/rif40_database_install.sh). There are whole sections in that version that are not in the GitHub one so presumably that version needs to be merged into master at some point.

• Version check: my version of PostgreSQL is 10.3, and the version checking in the script used string parsing which broke when the major level became greater than 9. I changed it to use a numeric version.
• I had to make some tweaks to the user martin, which I had created. It had also created a role of the same name, and that role had no password, which caused problems. Presumably I missed something out compared to running the Windows script, as I didn’t have that problem there.
• The SQL script sahsuland.sql is called as part of the run, and it ended with some errors regarding the role peter not existing. I ignored those for now.

Tweaks to Get the RIF to Run

I then had to manually revoke all elevated privileges for my user martin. I had created it as superuser, and that isn’t allowed for RIF users.

Manually created a schema called martin in the sahsuland database. Again, I must have missed something out in running the scripts.

Had to run the v4_0_alter-9.sql script, because of missing columns. But that was true on Windows, too.