3 ms·
You'll probably be happy to see the 1.17 dev version of gitea (which should be the next major release) actually has an artifact repository built in and it can e
by qbasic_forever 4y ago
You'll probably be happy to see the 1.17 dev version of gitea (which should be the next major release) actually has an artifact repository built in and it can even host docker images: https://docs.gitea.io/en-us/packages/overview/ https://docs.gitea.io/en-us/packages/overview/ You have to build the branch yourself right now though, despite being in the docs it's not in the release downloads yet. You might be able to get rid of nexus and just use this built in registry.
- KronisLV 4y agoThat's actually pretty cool to hear! Though a problem that I ran into with GitLab Registry was that sometimes the Docker image cleanup wasn't done properly and old layers kept lingering around and eating a lot of space. I sort of fixed it with some cleanup scripts and the occasional manual actions, but it still was a bit problematic since running out of space was a very real risk. Nexus does appear to have similar issues, should you use it as an intermediary for CI image builds (e.g. tag it with branch-date and tell your cluster to redeploy the image), but at least it has instructions for cleaning it up that work most of the time: https://help.sonatype.com/repomanager3/nexus-repository-administration/repository-management/cleanup-policies#CleanupPolicies-HowDockerRepositoriesUseSpace https://help.sonatype.com/repomanager3/nexus-repository-admi... Personally, I think that using the same software package for storing build artifacts and the source code itself is a bit risky, but maybe Gitea will be able to tackle those challenges with no significant issues!
- blktiger 4y agoIt probably depends on the scale of your builds, but I’d agree that it seems like a generally better idea to separate your docker image repositories from your source control.
- jonnycomputer 4y agoI just discovered that the latest gitea supports package repositories. I'd like to see something that works for R packages without too much hassle.