https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# k3s
a
This message was deleted.
l
Very broad question … have you tried anything? Read the docs?
b
yes tried, nothing works
any related docs ?
l
yeah the autoscaler docs
the autoscaler you’re trying to use … there’s different ones.
tried … nothing works … like what did you try … version of K3s and so on … hard to help with so little info
b
i tried with 1.26.5 version and disabled cloud controller manager and tried to setup aws cloud controller manager
i'm getting issues with provider id
Copy code
curl -sfL <https://get.k3s.io> | INSTALL_K3S_EXEC="server \
  --disable-cloud-controller \
  --disable servicelb \
  --disable traefik \
  --node-name="$(hostname -f)" \
  --kubelet-arg="cloud-provider=external" \
  --write-kubeconfig-mode=644" sh -
l
yeah the CCM can be a jungle … had to reverse engineer it pretty good to get a K3s cluster up and running in GCP
I don’t know AWS that well .. so. But, the logs in the CCM? Telling in any way? And look for potentially raising the verbosity of the CCM logs
b
Copy code
E1116 10:33:57.218536       1 static_autoscaler.go:337] Failed to get node infos for groups: wrong id: expected format aws:///<zone>/<name>, got <k3s://ip-172-31-89-21>
this is the issue
it's working
Thanks
l
What was the fix then?
b
the providerid was given by CCM
while installing k3s this is required to identify the node by CCM
Copy code
--node-name="$(hostname -f)"
l
Fair enough … thanks and glad that you where able to fix it.
🙌 1
Open in Slack
PreviousNext
Powered by