This message was deleted.
# rke2a
adamant-kite-43734
11/03/2023, 5:32 PMThis message was deleted.
c
creamy-rainbow-46562
11/03/2023, 5:42 PMwatching
creamy-rainbow-46562
11/08/2023, 5:39 PMHi @creamy-pencil-82913, sorry for the direct mantion, we are new on this Rancher world. And I´m seen you giving a lot of feedbacks, then I thought, he may can help us. 🙂
About our problem, is there any configuration to rancher watch the nodes and remove them if they are not healty?
c
creamy-pencil-82913
11/08/2023, 5:47 PMThis would normally be handled by the cloud provider running in that cluster. Are you deploying to AWS but not using the AWS cloud provider?
c
creamy-rainbow-46562
11/08/2023, 5:58 PMHum.. yes.
We have this ASG that manages the machine park. And in the "userdata" of the machine we have the script to start the agent and "join" to the cluster.
The addition of new nodes works pretty well.... the problem is, what to do with those "ghost" nodes after the ec2 dies.
c
creamy-pencil-82913
11/08/2023, 6:05 PMthe cloud provider will delete a node if the ec2 instance it is associated with doesn’t exist. that is specifically one of its jobs.
creamy-pencil-82913
11/08/2023, 6:06 PMc
creamy-rainbow-46562
11/08/2023, 6:07 PMhumm nice. I will take a look on that!
Thanks man, I will keep the thread updated
creamy-rainbow-46562
11/23/2023, 12:36 PMUpdate:
I did installed this provider and then I figured out that I also need the autoscaler available.
creamy-rainbow-46562
11/23/2023, 12:46 PMThe problem now is.... even with all pieces in place and looking good.... I'm facing now this messages:
Copy code
Failed to get node infos for groups: wrong id: expected format aws:///<zone>/<name>, got <rke2://ip-10-13-21-123And> as far as I could understand... We have to make some custom ProviderID, to attend this "aws" pattern...4 Views