https://rancher.com/ logo
Join Slack
Powered by
Hey So i'm trying to install a RKE2 cluster in my ...
# general
f
Hey So i'm trying to install a RKE2 cluster in my infrastructure i did follow all the steps in the documentation but i stuked in a problem that the container of kube-apiserver is never created , i have no clear error logs so i'm kind of stuck any idea pls ?
c
you sure you have enough resources?
if you see static pods not being created that’s about the only reason why
but also, its the second container created, after etcd. so… you’d need to be messing up something pretty fundamental if it’s not just resources.
f
I'm quite sure yes , my VM have 4 CPUS , and dynamic memoryù
i see etcd and kube-proxy pods the only one not creating is kube api
i did uninstall/install rke bin multiple time
The only error is see in the RKE service is : 
Pod for kube-apiserver not synced (pod sandbox not found), retrying
c
is etcd up yet? it won’t start the apiserver until etcd is healthy
you can also check the kubelet.log
f
Yes etcd and kube proxy are up and there is nothing wrong in their logs
i cant find kubeapi server log wich is normale cause there is no pod kubeapi server being created
and btw this is the docs i did follow : https://docs.rke2.io/install/quickstartù
the thing that confuse me is how i'm a supposed to debug this without like kube api logs or something
c
check to see that the apiserver static pod manifest has been dropped under /var/lib/rancher/k3s/agent/pod-manifests. if it has then check kubelet.log to see why it hasn’t started it.
if it hasn’t been created in that directory yet, check the rke2 logs in journald to see what it’s waiting for.
f
i already did check and the manifest of kube api is there
c
ok. so what does the kubelet log say
f
hmm didnt check to be fair
oh wait when you say kubelet log do you mean the service log ?
c
no i mean kubelet.log. there is no separate systemd service for the kubelet.
f
well no i didnt check it
c
why not
f
didnt think so , i'm gona check it tomorrow i dont have my working laptop anymore
hope i find something relevant
thanks btw
Hi @creamy-pencil-82913
So i did check the kubelet logs and nothing relevant
The first error is that the api is unavailable
ok wow did work
i did add some memory to the vm and works
idk if because i added ram or because i did restrat the vm
c
probably memory. that is why I asked first. if the kubelet doesn’t have enough resources to admit the static pod, nothing works.
how much memory did it have?
I assumed when you said it had “dynamic memory” the baseline allocation was at least the minimum required in our docs
f
i did give it 4096mo now as minimal memory + dynamic
5 Views
Open in Slack
PreviousNext
Powered by