This message was deleted.
# general
a
This message was deleted.
b
that's certainly the direction suse is pushing toward. depending on your environment, harvester might not quite be ready for prime time though. also right now they don't recommend hosting rancher on top of harvester, so you'd need a separate environment for that.
at the moment i'm cheating and running libvirt on the bare metal and using vms for both harvester and a separate k3s cluster to run rancher.
s
Okay, we could run Rancher on a separate cluster. I am mostly afraid of "magic stuff" that Rancher and Harvester cloud do and potentially break things. With our current solution based on rke2, I am confident enough to run in production.
b
the main thing keeping me from actively recommending harvester for real production use is the lack of zero-downtime upgrades.
https://blog.stevecoinc.com/2023/05/harvester-hci.html if you're at all interested in a short version of my experience with harvester. ๐Ÿ™‚
s
Sure, was looking for some blog posts about experiences. ๐Ÿ˜„
Also my first few minutes using Harvester are not very promoting for Harvester. ๐Ÿ˜„
What do you think about Rancher? We cloud also implement Rancher without Harvester in our setup.
b
the k8s expert on my team is a huge fan, which is enough for me. ๐Ÿ˜†
s
What kind of workload do u run on Rancher? We mostly run customer applications, and also implementing some kind of public cloud.
b
i'm still trying to figure stuff out personally. my coworker who recommended rancher ran everything possible on k8s managed by rancher.