[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
[rdf4j-dev] dockerhub deployments
|
Hi Bart, everyone,
Just wanted to let you know, first of all, that I finally got around to getting myself sufficient permissions for our dockerhub repo.
Couple of questions: First of all, I notice the README on Dockerhub is a little out of date, is that content fed from anywhere in our git repo, or is written manually/separately on Dockerhub itself (in other words: where should I update)?
Our current docker images are manually pushed after a release. I am not terribly familiar with the workings of Dockerhub but it appears to have an automated build feature that can inspect our github repo and trigger on particular tags. Is that worth exploring a bit further?
Another thing: we produce an AMD and an ARM image. Is there really value in having both? I mean I'm not against it but I kinda suspect that the vast majority of our users are on AMD architectures.
Even if we continue releasing an ARM image, I'm thinking that our default should just be the AMD one, so I'd suggest that we use tag names like this:
- rdf4j-workbench:latest <-- latest AMD image
- rdf4j-worbkench:3.4.3 <- AMD image for release 3.4.3
- rdf4j-workbench:arm64v8-latest <-- latest ARM image
- rdf4j-workbench:arm64v8-3.4.3 <-- latest ARM image
Finally: anything else I should be aware of in terms of Dockerhub, and where/when/how we produce docker images?
I suggest that we make it a habit to notify if we intend to deploy a new docker image following a release, so that we don't tread on each other's feet. I'll do a build for the 3.4.4 one if you don't mind, I need some hands-on practice :) I'll keep using your tag naming scheme for now, not the one I just proposed.
Cheers,
Jeen
