Hi,
Is there a recommended deployment strategy for openBIS that separates dev and prod deployments? A couple of scenarios:
- Users want me to add a new field to an existing SAMPLE object, e.g. “Date discarded“. This is a small change with likely zero undesirable side effects, but I still hate pushing this change straight to prod.
- Less trivial: I have to upgrade openBIS itself, or postgres from v11 to v15, or the Ubuntu OS to a newer LTS version. Ideally I want to test out the upgrade in a dev environment first.
Our OpenBIS deployment is on an internal VM. I’ve tried simply cloning the VM before, but this usually results in problems as lots of ip address settings get duplicated, and the front end is unable to resolve which VM to point to. I imagine that this will be the same if I were, say, running OpenBIS on a cloud VM on the same cloud network as well.
I was hoping that there’d be some recommended strategy of how to run 2 or more OpenBIS instances in parallel at all times, where dev always syncs to prod on a scheduled basis? Perhaps I pg_backup the prod VM, then pg_restore to dev VM every day or something?
Regards,
don