Split relay roles per server type #116
Conversation
|
This is for #117 |
| hosts: | ||
| olympics2024.nos.social: | ||
| relay_image_tag: olympics_latest | ||
| dev-olympics2024.nos.social: |
There was a problem hiding this comment.
I think we can nix dev-olympics2024.nos.social, since we are trying to move to personal dev boxes rather than having one for every service. Is there an actual droplet here that we should tear down?
There was a problem hiding this comment.
tbh I feel weird about putting the olympics2024.nos.social relay in this inventory file too, because it wasn't deployed with the nostr-rs-relay playbook and tbh I'm not totally sure what would happen if we ran it on the live box. Maybe we can leave the vars in here but leave hosts empty with a comment that we don't have any production servers using this role right now?
There was a problem hiding this comment.
I'm ok with that. I didn't know we wanted personal servers but that's also fine. I had asked @gergelypolonkai to create a dev-olympics2024.nos.social box but we can just change the name to daniel.dev.nos.social instead. I assume the idea is reusing those for every type of test on other unrelated projects.
There was a problem hiding this comment.
I only added my personal host for the new inv and left olympics as before so the old inv and role works as it was. I have a relay deployed at my test ip but it still needs to have a dns name for the ssl to work, I think @gergelypolonkai is working on that.
There was a problem hiding this comment.
Oh yeah sorry I didn't communicate on the personal dev boxes. We were basically doubling our DO costs by having a dev version of each service and we really only used them to test ansible scripts.
roles/relay/files/strfry.conf
Outdated
There was a problem hiding this comment.
🤔 was this not being used? Same question for allowed_rules.js.
There was a problem hiding this comment.
Exactly, it's living solely in the repo here, much easier to update.
Co-authored-by: mplorentz <[email protected]>
We used to maintain a single relay and had to reuse the setup for the nostr-rs-relay while working under the Olympics deadline. Now that the new relay is live, we need to make some changes for better maintainability.
This refactor introduces separate inventories and roles for strfry relays and nostr-rs-relay relays. Although they are similar, there are differences in directories and Docker images. In future work, more differences may appear, so this structure allows us to manage those changes for each type of relay more cleanly.
I also removed unused config settings, which are now handled through their corresponding Docker repositories' automated deployments.