Time/Place

Time: 12:00 pm Eastern Time (US)

Zoom Link:  https://lyrasis.zoom.us/j/396037251

Attendees

1. Danny Bernstein 
2. David Wilcox 
3. Demian Katz 
4. Geoff Scholl
5. Jared Whiklo 
6. Mersed Kahrimanovic
7. Peter Eichman 
8. ?

Objectives

1. Establish consensus on desired Camel Tool Box deployment options
2. Establish major and roles for development and testing
3. Establish timeframe for development

Reference

• https://github.com/fcrepo-exts/fcrepo-camel
• https://github.com/fcrepo-exts/fcrepo-camel-toolbox
• https://github.com/umd-lib/umd-fcrepo-docker
• https://github.com/umd-lib/umd-camel-processors
• https://github.com/umd-lib/umd-fcrepo-docker/blob/develop/activemq/pom.xml

Agenda

1. Introductions
2. What are the possible approaches?
   1. Peter Eichman on UMD's setup
      1. What worked, what didn't, recommendation
   2. Other approaches
3. Which one is best?
4. What is the targeted deadline for this work?
5. Who, what and when?

Notes

Peter Eichman  explained his approach.

He is using the camel container provided by ActiveMQ which has allowed them to eliminate internal queues and reduce complexity.  It is not clear to me (Danny Bernstein ) if running in ActiveMQ's jvm makes this possible or if he's made assumptions about deployment (namely all services are being deployed together) that allowed them to eliminate some of these queues.

To have the best of both worlds,  we could (according to Ben Pennell ) make the route endpoints configurable which would allow deployments to make use of direct messaging.

Proposal 2: 

• 1 code repository with multiple modules. 
• Each microservice in a module.  A spring boot app (module within the same repo) pulls all the services together in one application. 
• Configuration files can define which micro services you want to run as well as the endpoint configuration for the routes.
• It would allow folks to run all services in one docker container or each service in a separate docker container.  

Proposal 3: Same as above but each micro service lives in its own github repository and is versioned separately. The application that pulls them all together also lives in a separately versioned repo.

The consensus seemed to be inclining towards Proposal 2.

Demian Katz and Geoff are game for testing.

Mersed also willing to help with testing and development.

A sprint sometime in April or early May would be good.

Action Items