kubeadm v1.13 ignores --kubeconfig-dir on phases

[tommyknows]

Is this a BUG REPORT or FEATURE REQUEST?

BUG REPORT (?)

Versions

kubeadm version (use kubeadm version):

kubeadm version: &version.Info{Major:"1", Minor:"13", GitVersion:"v1.13.0", GitCommit:"ddf47ac13c1a9483ea035a79cd7c10005ff21a6d", GitTreeState:"clean", BuildDate:"2018-12-03T21:02:01Z", GoVersion:"go1.11.2", Compiler:"gc", Platform:"linux/amd64"}

What happened?

I'm writing a script to automate kubernetes deployments with kubeadm. Up until now, I have used the following command to do all the steps regarding bootstrap-tokens on a remote server:

kubeadm init phase bootstrap-token all --config "$kubeadm_config_file" --kubeconfig "$kubeconfig_file"

With kubeadm v1.13, this does not work anymore - there's no --kubeconfig flag anymore.
I'm now trying to use

kubeadm phase bootstrap-token all --config "$kubeadm_config_file" --kubeconfig-dir "$kubeconfig_dir

This does not work:

can not mix '--config' with arguments [kubeconfig-dir]

I have removed the --config flag (as this may not be relevant in my case), and set $kubeconfig_dir to a path that contains the admin.conf file or giving the path to a valid kubeconfig file directly into this variable.
However, I am always getting the following output:

error execution phase bootstrap-token: failed to load admin kubeconfig: open etc/kubernetes/admin.conf: no such file or directory

What you expected to happen?

I expect kubeadm to either read from the kubeconfig-file (supplied by --kubeconfig-dir), or search the --kubeconfig-dir for a admin.conf file.

How to reproduce it (as minimally and precisely as possible)?

• have a running kubernetes master
• copy the kubeconfig file (admin.conf) to some path
• run `kubeadm init phase bootstrap-token --kubeconfig-dir [KUBECONFIG-PATH]

[yagonobre]

Hi @tommyknows can you post the output of echo $kubeconfig_file?

[yagonobre]

/kind bug
/priority awaiting-more-evidence

[tommyknows]

Well, it's a temp-directory (/tmp/tmp.[something]), created by mktemp.
For example:

ls /tmp/tmp.rTQrPIMhog/
admin.conf  controller-manager.conf scheduler.conf

I've also tried setting $kubeconfig_file to /tmp/tmp.rTQrPIMhog/admin.conf, that doesn't work either.

[yagonobre]

@tommyknows I'll send a pr with the fix soon
/remove-priority awaiting-more-evidence
/priority important-soon

[k8s-ci-robot]

@yagonobre: Those labels are not set on the issue: priority/

In response to this:

@tommyknows I'll send a pr with the fix sson
/remove-priority awaiting-more-evidence
/priority important-soon

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[k8s-ci-robot]

@yagonobre: Those labels are not set on the issue: priority/, priority/awaiting-more-evidence

In response to this:

@tommyknows I'll send a pr with the fix soon
/remove-priority awaiting-more-evidence
/priority important-soon

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[yagonobre]

/assign