[RAVE-382] Consolidate and move all person data initialization to be done by rave-portal

• Type: Sub-task
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 0.6-INCUBATING
• Fix Version/s: 0.7-INCUBATING
• Component/s: rave-core, rave-portal, rave-shindig
• Assignee: Ate Douma
• Reporter: Ate Douma
• Created: 07/Dec/11 09:06
• Updated: 23/Feb/12 14:41
• Resolved: 07/Dec/11 10:07

Description

Currently both rave-shindig and rave-portal seed the person (and related) data during initialization.
As these data sets overlap (erroneously) and the the initialization order of these two web applications isn't guaranteed, this can cause an initialization failure (duplicate key error) at startup.

If this happens, a temporary workaround for the 0.6-incubating release binary demo is:
1) first make sure to remove the (broken) rave database, like with: $ rm /tmp/rave_db.h2.db
2) before first start of tomcat after extracting the binary demo, temporarily move the webapps/ROOT.war file outside the webapps folder so it won't get deployed
3) now start tomcat for the first time, and once fully started stop it immediately again (don't try to access the portal yet)
4) move or copy back the ROOT.war back into the webapps folder
5) now tomcat can be started normally and be used for the demo
6) whenever you want to 'reset' the database again, rewind back to step 1)

The proper solution for this problem is consolidating and moving all person data (and for related objects as groups and their associations) from rave-shindig initial_data.sql to the rave-portal (resources) initial_data.sql
This also is needed (manually) if you want to build and run Rave from source, including for the 0.6-incubating release/tag, when using the cargo:start procedure as that automatically managed the deployment of the wars (above manual steps cannot be performed then).

I'll create a patch for this change to be applied to the trunk, but which also can be applied when using the 0.6-incubating release/tag sources.