This message was deleted.
# generala
adamant-kite-43734
08/22/2023, 4:40 PMThis message was deleted.
a
acoustic-addition-45641
08/22/2023, 5:06 PMHaven't tried it personally, but is this what you are looking for?
https://ranchermanager.docs.rancher.com/pages-for-subheaders/install-cluster-autoscaler
a
abundant-hair-58573
08/23/2023, 1:17 AMWe're using the cluster-autoscaler in Rancher with an AWS autoscale group. It works great for scaling out, but when scaling in it leaves the terminated nodes in Rancher, we clean them up manually every day. I keep meaning to get back to figuring out the scale in portion but for now manually removing the nodes takes about 20 seconds every morning
s
stocky-fall-82040
08/23/2023, 12:06 PMWe create a node termination script with the user data that curls rancher for a node_id and then runs a curl deleteon that node id. Both the node_id retrieval and delete call the API and used bearer token. We then setup a node-termination systemd script, that executes the node termination script, which gets executed when the node is shutdown/terminated.
stocky-fall-82040
08/23/2023, 12:08 PMZane, we do use Cluster Autoscaler with AWS ASGs. When nodes join the cluster, since we use node groups, it always shows as in the UI as "Not In A group". The nodes in he ASG are all worker nodes. I wish it would join an existing node group instead of it being its own separate thing or wish it could be named appropriately.
👍 1
a
abundant-hair-58573
08/23/2023, 3:03 PMAh, so you have your nodes basically deleting themselves from Rancher in their own shutdown script? I tried something like that but the script wasn't executing when the ASG killed the instance
5 Views