I'm in the process of automating — or at least attempting to — the installation of several Atlassian products we've bought. Both Jira and Confluence seem to support unattended installation, to one degree or another (Jira's seems more fully-baked than Confluence's). I'm not finding any similar documented feature for Crucible. It seems like Crucible is "add run-user and firewall exceptions; Download and dearchive the ZIP; fire the application up; log in and configure it." Included in that "and configure it" being the setting up of the DB connection.
While Confluence has a somewhat similar limitation, Confluence at least doesn't default to starting to run against an internal database that you have to migrate if you want to use an externalized DB.
Crucible appears to require starting from that internal database and then doing a destructive migration to the externalized DB ...rather than just "configure your JDBC string and only destroy content if you explicitly say so". Is this actually the case or am I missing something somewhere. If this install limitation is as it seems, it really kind of bones me on being able to automate redeployments (e.g., for use with AWS's AutoScaling feature).
Any pointers on how to get around this apparent limitation?