Community Help

[suse-coder]

Hi,
I have developed this helm chart, which is fully working and has all in it (decomposedfs, s3, posix, rwx, clamav, openldap):
https://github.com/suse-coder/helm

It could be good if there is more cowork as upgrading each version (and opencode is fast in its cycles) is quite a work. If you like you can merge my work (as i am the original creator of the opencloud helm charts). As al these tests over time are quite a lot of work I would love if you can help to get a fully working helm chart for all.

[suse-coder]

I like that you only use one opencloud pod. That is a good decision. What you should add is decomposedfs, rwx (for cephfs) and clamav and openldap (or its successor). And have rolling updates (not recreate) and multiple pods (at least two with rwx for cephfs)

[suse-coder]

I love that you use talos (and i guess cilium + cilium gateway api + fluxcd operator)

[suse-coder]

Will also update to new https://github.com/osixia/helm-charts for openldap (multi opencloud instances) and use helmfiles for keycloak, clamav and openldap

[Tim-herbie]

Hi @suse-coder,

Thanks for reaching out. I’d definitely be interested in avoiding duplicated effort and working together on the OpenCloud Helm chart.

I try to keep dependencies as minimal as possible. I’m not sure whether we should integrate additional components like OpenLDAP. However, I like your thoughts on enabling the possibility to run OpenCloud in HA mode. What would you integrate into the helm chart here?

I’ve already set up a working HA OpenCloud deployment with NATS, OpenLDAP, and Collabora, which worked well. But I’m not sure how to integrate this into the Helm chart without increasing complexity for single-node users who want a streamlined setup. Maybe it would make sense to keep the base components clean and provide an “HA OpenCloud” example separately.

And yes, I use Cilium as my CNI with Cilium gateways and Flux to manage everything within my RKE2 cluster. I switched my homelab setup a few weeks ago from a 6 node to a single-node cluster to save some power.

[suse-coder]

I already did today all the work will do a PR just now. Works fully now with rwx, clamav, openldap etc. Just added the best stuff you dont have, everything else stays from your code as i like the lean thing. We also have opencloud in production for half a year so added the stuff that improves it. I moved keycloak out so all is professionally in a helmfile.

[suse-coder]

You should switch to talos. It is really much worth it.