Description of problem: Dynflow Worker process is now part of the executor, meaning they can't be restarted independently. As a first step to improve scalability, following two stories will be implemented 1. As a satellite user doing an install or an upgrade, the installer should create both the orchestration and worker processes with sane defaults. Both processes will run on the same machine as the satellite. 2. As a satellite admin, I should be able to restart a worker process without impacting the uptime of the satellite or losing data. Verification steps will be added later when we know names of new services.
Verified with Sat 6.8 snap 7.0. Sidekiq is now used. It's now possible to: - change a number of workers, number of their threads, their queues - view log using "journalctl -u 'dynflow-sidekiq@*'" - restart a worker without getting to inconsistent state Note that this intentionally does not apply to external capsules. See this demo: https://www.youtube.com/watch?v=8xomzg_FBIs&feature=youtu.be&t=485
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: Satellite 6.8 release), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2020:4366