11.1 linkerd fails during install

chadwitk

I've seen this exact same question a few times, but no resolutions.
I'm running ubuntu on aws, and up to this point everything has gone as expected.

Any suggestions?

chrispokorni

Perhaps typos on the name regex ... line @chadwitk

-Chris

chrispokorni

Hi @chadwitk,

I see many timeouts to various port numbers. Is your SG open, per the AWS setup video from the introductory chapter?

Regards,
-Chris

chadwitk

wide open in and out:

chrispokorni

Hi @chadwitk,

What Ubuntu release is running your VMs?
Did you get any errors during the CRD or the linkerd installations?
What's the state of your cluster overall?

kubectl get nodes -owide

kubectl get pods -A -owide

Does top on either node show any processes consuming too much CPU or memory?

Regards,
-Chris

chadwitk

@chrispokorni thanks for helping!
runnikube Ubuntu 20.04.6 LTS
crd install:

fullinstall:

nodes:

pods:

any attempt to 'top' just gives 'error: Metrics API not available'

chrispokorni

Hi @chadwitk,

I am concerned about the coredns pods. Did they ever work (were steps 3.3.4 and 3.3.5 successful)? Or did they stop working around lab steps 9.3.8 and 9.3.12 after Corefile edits?

Please delete both coredns pods from the kube-system namespace and see if they are restarted successfully.

In the meantime, what is the output of

kubectl -n kube-system get configmaps coredns -oyaml

The metrics API error is typical for the kubectl top with no metrics-server installed or initialized.
What about simple top?

Regards,
-Chris

chadwitk

I didn't have any issues in section 3
The coredns pods failed to properly re-start after deleting (both re started with immediate crash loop backoff)

Thanks again for the help, I'm going to go back to take another look at 9.3.* and see if there is something in there I got wrong

chadwitk

@chrispokorni thank you again for the help, that was the problem!!