https://rancher.com/ logo
Title
b

bored-rain-98291

07/21/2022, 1:39 PM
Good morning all. RKE2 seems to be a bit different than im used to. We have it running in a vmware cluster aka bare(ish) metal. Everything ive deployed has had issues but i think its primarily due to being metal and not cloud. I seem to have most of it working correctly except a few pods: helm-install-rke2-core-dns-xxxx and helm-install-rke2-canal-xxxx. Not sure where these came from, Anyone know what these are? many thanks
r

rapid-helmet-86074

07/21/2022, 2:47 PM
On a normal successful RKE2 install (on VMs, just KVM on CentOS rather than VMware) I get a helm-install-rke2-coredns-, helm-install-rke2-canal-, helm-install-rke2-ingress-nginx-, & helm-install-rke2-metrics-server- . That's held for v1.21 & v1.22, though the Rancher docs have a different list than I get so some older verison is somewhat different (though I think these were the same). My understanding of the point of these is that things like kubelet, kube-apiserver, kube-proxy, kube-scheduler, kube-controller-manager, etcd (& maybe a few more I'm forgetting) get run as child processes to rke2-server service or as static pods with the RKE2 install. However, there are additional pieces to install to get things working, like the CNI, so those are installed to the fresh Kubernetes cluster (rather than static pods) with a helm job and that's what those are.
If you look at the crictl ps command on https://gist.github.com/superseb/3b78f47989e0dbc1295486c186e944bf you can see your static pods.
b

bored-rain-98291

07/21/2022, 2:48 PM
I use to work with another guy with the same name many years ago. Did you ever work for Weblogic?
I worked at PricewaterhouseCoopers and we had a weblogic consultant with the same name 🙂
thanks for the info!
👍 1
r

rapid-helmet-86074

07/21/2022, 2:49 PM
Nope, but my understanding of my last name is it's Welsh for "cottager", which I'm assuming = "serf", so there are pockets of us scattered around.