Can't access cluster from kubectl after Rancher certificate update

Rancher 2.x
1

Hello,

I am using Rancher 2.8. I replaced the default certificate with a new one following the “Updating from a Private CA Certificate to a Public CA Certificate” procedure from Updating the Rancher Certificate | Rancher

I used the full certificate chain, but after updating the certificate I cannot access the underlying cluster even when I replace the KubeConfig file. The message I get is:
“Unable to connect to the server: tls: failed to verify certificate: x509: certificate signed by unknown authority”

I ran superseb/ranchercheck script and it reports:
Certificate chain is complete, connection to https://rancher.example.com established successfully.

Please help or point me in a new direction, I’m going around in circles.
Thank you.

1 Like
2

Just asking, have you imported the RootCA to your machines?

3

Hello @wioxjk thank you for answering.

I haven’t imported RootCA to any machines. Would this be the local cluster macine (i.e. Rancher manager) or the remote cluster nodes? Why is only RootCA needed in that case?

I only did force update from Continuous Delivery as indicated by the procedure. Could you send me the procedure to import the certificates to the nodes that need to be updated?

4

Depending on the machine you are running kubectl from, you might want to look into different way of importing your RootCA into that one.

https://ubuntu.com/server/docs/security-trust-store

https://www.eduhk.hk/ocio/content/faq-how-add-root-certificate-mac-os-x

5

Anyway, you probably need to import the root-cert into rancher aswell:

6

I tried adding the certificate to the client machine, this did not change the error.
I’m having trouble with this one: About Custom CA Root Certificates | Rancher
My Rancher manager was not installed with docker, I’m assuming there should be an equivalent of this action for helm chart, but I was not able to find it. Could you help with this?

Thank you :slight_smile:

7

Did you check this?

8

Hello, I have: remote cluster becomes unavailable if I do this. I tried a bunch of other posts related to this, but I still did not resolve the problem.