Thinking ahead, orchestrating Armadietto containers

Just thinking out loud… I haven’t really done much yet with the code in 2022 as I’m still fixing some stuff in another project after some comments I read here on the forums :).

Once we have Armadietto containerized we’ll want to orchestrate.

I wonder what that looks like. Most of my thinking is around kubernetes.

Persistent volume claims (e.g. shared file system) shared by a deployment of Armadietto nodes behind a load-balancer might run into sync issues.

So maybe the way forward is a shared redis persisted to a PVC, since it’s the one datastore already integrated: although it looks like redis is still experimental in Armadietto? I haven’t yet tried it.

Either approach should provide for sufficient metrics and monitors in most deployments (:crossed_fingers:). I don’t think we need prometheus integration or some such to make it viable.

So, in my head, if redis integration is actually up to the job already, no further code changes needed to bring up clusters? Just devops?

It looks to me like Redis support was removed (or at least not ported) in the fork from reStore to Armadietto:

Maybe @les could update us on the status of that?

Oh I see. I saw it mentioned in the tests. Good to know, thanks.

So redis has a spec file but it’s also mentioned several times in the README.

The branch feature/redis hasn’t been worked on in four years and I think @les has moved on from this project (judging by comment in another PR).

I also see @DougReeder, you have a branch to stress test Armadietto :+1:.

Maybe we can be data driven here and see if it’s worth putting some time into redis or to just rip it out of the README not to confuse future generations.