shwstppr commented on issue #7504: URL: https://github.com/apache/cloudstack/issues/7504#issuecomment-1541996916
@tuanhoangth1603 This is how cluster and scaling work: - Cluster is deployed with `n` size. CloudStack will deploy 1 (default) control node and n worker nodes. For the public IP it will create a firewall rule to open ports 2222 to 2222+n. It will also set up port-forwarding for 2222+x to port 22 for each of the node VMs. (2222 will be forwarded to the control node, 2223 to the 1st worker node, 2224 to the second, and so on...). It will also open port 6443 and add a corresponding port-forwarding rule for the control node on the same port. - When you scale UP, CloudStack will deploy new required worker node VMs, old firewall is deleted and a new firewall rule is created for 2222 to 2222+newsize. Port forwarding rules are set up as explained in the first point. Now, for you, ACS failed to delete and create the new firewall rule (I'm not sure why but it could be the additional instance that you mentioned). If you can give the output of listKubernetesClusters API, screenshots of the cluster instances list, firewall rules, and portwarding rules for the cluster public IP we can have a deeper look. As you say, k8s part is working fine. So what I suggested doing is, - make sure firewall rules are provisioned correctly, ie, port 6443 and 2222 to 2222+clustersize - make sure port-forwarding is provisioned correctly as explained above in point 1 - Wait for KubernetesClusterStatusScanner for 5 mins. It runs every 5 mins. There is no way to manually run it. And it should fix the cluster state. This shouldn't require stop-starting the cluster -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
