I’m not sure – but from my POV – our community isn’t that big. We can’t afford to host everybody’s modules in our nexus repo.
Yup, see all the posts aboved.
We need to figure out how to get this down to a reasonable size. I’d rather not dedicate a whole server to JUST Nexus/mavenrepo – that seems like a waste of resources.
I did some investigation and a good chunk of the space was taken up by the trash directory for snapshots – which should have been deleted – but never was. I cleaned that up and now it’s only 16GB now. This is a size I’m happy with. No changes are immediately necessary now.
Perfect, thanks.
So lesson learned here, is to perhaps troubleshoot circumstances before labeling them as critical issues or proposing fundamental changes to infra… 
Nothing to see here, folks… move along. 
Thanks everyone.
It’s still something we need to do and it was critical. I guess I didn’t think to drill deeper to see where the larger files were…
No so fast, This is something we still need to consider. We just bought a little bit of time. This will still cause issues. I’ll be monitoring Nexus’ disk space usage closely.
One thing I’m a little confused about is whether us module developers should be deploying to the OpenMRS nexus repo. Way back when we got started releasing the OpenHMIS modules 4 years ago I thought that the standard practice was to deploy our modules to the omrs repo (perhaps I was mistaken). It seems like there is value in have a single, community-wide, repository for our modules, if that is still a viable option.
At this point I just want to know what we should be doing.
It shouldn’t be a problem for established projects as ours.
You can configure repos in module poms or globally for your system in settings.xml. If you are using the SDK, it configures repos for your system when you call setup-sdk.
Yes, our Maven Repo can be easily made read-only once we are ready for the switch.
Anyways, now after emptying the trash and disabling proxy for maven central we can migrate at our own pace. Thanks @r0bby!
@ibewes, it’s unquestionable it helps to have a single repo for OpenMRS modules. Currently it’s mavenrepo.openmrs.org and modules.openmrs.org, but as we grow I’d like to make it easier for people and publish all artifacts to Bintray under openmrs org at Service End for Bintray, JCenter, GoCenter, and ChartCenter | JFrog
Rafal answered this, but to be more direct: keep doing what you’ve been doing. We still want module developers to deploy to the OpenMRS maven repository.
(At some point in the medium term, we’ll shift to using bintray, but you’ll be amply informed about what to do when this happens.)
All good news, thanks @raff!
Not, really. I can’t say we’ll hang onto everything forever…
Hey guys, just to let you know, we are about to sponsor a full Artifactory Cloud instance for selected OSS projects, not only an account on oss.jfrog.org (that’s for SNAPSHOTs and development process). We are finalizing the process now, but if you want to start today, you can go ahead and start a trial and I’ll make sure to convert it to the free account once we have the process finalized and your project approved.
@jbaruch, that sounds awesome! I think we would definitely be interested in this, but I also think it might take us some time to think about shifting from our self-hosted nexus to there.
No prob at all, take your time. I am, of course, more than happy to answer any questions you have and help with the migration.
@jbaruch, thank you! Do you have a hard limit on how much storage we get?
I’m signing us up and gonna look at it. Worst case we can torpedo it.
I set up openmrs.jfrog.io with the email of infrastructure AT openmrs DOT org.
And we’re up – that was fast! Provisioned on AWS – pretty quickly!
@r0bby, could you please share a bit more details about openmrs.jfrog.io?
Wow, you’re fast. Keep that trial open for now, it’s fully operational, so go ahead and put it to use! We’re working on the procedure of migrating it to a permanent opensource free server. Please let me know (by email to jbaruch@jfrog.com) when it is about to expire so I’ll know to extend the trial if we won’t be ready in a month.
Re the cap, we’ll do our best to accommodate your storage and bandwidth requirements (in sane limits, of course 
) . Can you please give me an estimate? I understood you have a lot of snapshots that you don’t really need, will having the last 8 be enough? If so, we’ll establish a cleanup for snapshots, so take that into account for your estimate.
Thanks!
Emptying the trash of old snapshots…it’s about 11-13GB total
The old snapshots are 11-13GB or what’s left is 11-13 GB?
I’m no longer on the infra team. Hopefully others can help out with this. But yeh – total our Nexus Repo is about that size. It DID bloat to almost 70GB at one point…if the old snapshots aren’t purged from the trash directory…it tends to grow at an insane rate.