communishift is the name for the OpenShift community cluster run by the Fedora project.
It's intended to be a place where community members can test/deploy/run things that are of benifit to the community at a lower SLE (Service Level Expectation) than services directly run and supported by infrastructure, additionally doing so in a self service manner.
It's also an incubator for applications that may sommeday be more fully supported once they prove their worth.
Finally, it's a place for Infrastructure folks to learn and test and discover openshift in a less constrained setting than our production clusters.
- Running OpenShift 4.1
- 8 compute nodes and 3 master nodes that are all dell fx1 blades
- Total vCPU: 552 vCPU
- Total Memory: 1.72 TiB
For now, Users in the group 'communishift' will be granted access via Fedora ipsilon and have a quota of 5 projects, 10pods and 5 volume claims. If you need additional resources, please file a ticket. Make sure to note what you need them for.
We hope to open access to larger groups soon after a trial period.
We reserve the right to remove at any time, any access/application/content for any reason.
- It is 100% up to the community member or group to write/maintain/deploy/backup/manage their app.
(but see best practices below)
- Fedora Infrastructure will keep the cluster up and functioning, anything above that is up to YOU.
- You can maintain your application however you like, but here's some best practices:
- Create a pagure/github/gitlab project for your application and use its git to track your config. Allow people to file tickets and PR's against it.
- Make sure your appication has a clear note where to file issues or how to contact the owner(s)
- backup your application config and data often. (see 'oc export')
Questions and Answers
- How does this relate to the Fedora Infrastructure Private Cloud?
- It replaces it. We found it difficult to deploy and manage openstack with the resources available to us, so we elected to switch over to OpenShift.
- My app/think is super popular and I want it to be more supported, what can I do?
- Submit a RFR asking infrastructure to take it on and move it to a higher SLE place/state.
- The cluster was reinstalled and I lost my work, what can I do?
- You should have backed up your application and data as we asked you to. Sorry.
- I need more (cpu/storage/apps/privs), what can I do?
- Ask for them in a ticket.