How can I automate Crowd installation and configuration for deployment on a Docker cluster.

Question

I'm using the blacklabelops/crowd Docker image as a baseline and auto-generating the crowd.cfg.xml file in the ENTRYPOINT script for my image so it contains our server ID, license, and database settings.

I'm having trouble figuring out how to initialize the database though -- when Crowd starts up, it doesn't create the schema, instead generating errors in the DB logs because the requested tables don't exist.

Is there a way to specify all relevant parameters at startup so Crowd will initialize if settings don't exist and use existing settings if the database is found? Or do I have to run a bootstrap container and manually configure Crowd before I can launch my runtime containers for an HA solution?

daniel.x.diaz@capgroup.com · Answer

any update on this question?

Marcin Kempa · Answer

Hi @Gordon Shankman,

Indeed there is no headless setup for Crowd at the moment. However, the approach you've taken with generating crowd.cfg.xml, looks like a way to go.

In order to overcome the problem with empty database, I would suggest doing a backup and then reapplying it on a fresh DB before Crowd is started. It can be a DB dump from a previous startup as Crowd will automatically upgrade DB schema and data to a version you currently run.

You've mentioned HA solution you are building with docker containers. I guess it is more disaster recovery mode then active-active clustering?

Did you tried our Crowd Data Center for HA capabilities in Active - Active cluster mode?

Hope that helps,

Marcin Kempa