Roberto Carratalá bio photo

Roberto Carratalá

Linux Geek. Devops & Kubernetes enthusiast. SSA @ Red Hat.

LinkedIn Github

You missed your kubeconfig file of your OpenShift cluster? Your dog ate your kubeconfig file? No worries! Let’s regenerate it in a easy and automated way!

Let’s dig in!


A file that is used to configure access to clusters is called a kubeconfig file. This is a generic way of referring to configuration files. It does not mean that there is a file named kubeconfig.

When an openshift 4 cluster is generated a Kubeconfig is created with a system:admin user and with administrator access privileges to the cluster.

Usually the system:admin user kubeconfig is available at:


But sometimes the folder is not accesible, because it’s located in another host, or by mistake was deleted.

So… how can you regenerate your kubeconfig file and set client certificate authentication to system:admin?

NOTE: for obvious security reasons the kubeconfig file for system:admin only can be regenerated by a cluster-admin user, so please log in with your cluster admin privileged user to do this procedure.

Generating the Kubeconfig File for system:admin user

1. Export Useful Variables

Define the Authentication Name for your new key and the new kubeconfig name:


2. Create a Certificate Request for system:admin user

Let’s start to generate a Certificate Request for the system:admin user.

A CertificateSigningRequest (CSR) resource is used to request that a certificate be signed by a denoted signer, after which the request may be approved or denied before finally being signed.

In our case we want to request a CSR for the system:admin user and we need also to provide the CommonName as system:admin and the OrganizationName as masters:

$ openssl req -new -newkey rsa:4096 -nodes -keyout $AUTH_NAME.key -out $AUTH_NAME.csr -subj "/CN=system:admin/O=system:masters"
Generating a 4096 bit RSA private key
writing new private key to 'auth2kube.key'

3. Create a signing request resource definition

Once we have the csr file generated from the earlier step we need to create a signing request resource definition for request a certificate for our new kubeconfig file:

$ cat << EOF >> $AUTH_NAME-csr.yaml
kind: CertificateSigningRequest
  name: $AUTH_NAME-access
  - system:authenticated
  request: $(cat $AUTH_NAME.csr | base64 | tr -d '\n')
  - client auth
$ oc create -f auth2kube-csr.yaml created

as you notice the usages are defined as client authentication and the request is passing the content of the csr file.

You have more information about this procedure and for the CSRs in the AuthN-AuthZ Kubernetes section

4. Approve the CSR and extract the client certificate

Now the Certificate Signing Request is waiting to be approved or denied. Let’s approve them!

$ oc get csr
NAME               AGE   REQUESTOR                                            CONDITION
auth2kube-access   1s    admin                                                Pending
dev-aws-926rz      41m   system:serviceaccount:dev-aws:dev-aws-bootstrap-sa   Approved,Issued

$ oc adm certificate approve auth2kube-access approved

And now we can extract the client certificate from the csr itself:

$ oc get csr $AUTH_NAME-access -o jsonpath='{.status.certificate}' | base64 -d > $AUTH_NAME-access.crt

5. Add the system:admin user entry into the new kubeconfig

Now that we have in our localhost our certificate let’s use them. But how can we use our certificate within our new kubeconfig file?

We need to define a new entry in the kubeconfig file for the system:admin using the client-certificate and the client key that we obtained and generated in the earlier steps:

$ oc config set-credentials system:admin --client-certificate=$AUTH_NAME-access.crt --client-key=$AUTH_NAME.key --embed-certs --kubeconfig=/tmp/$NEW_KUBECONFIG
User "system:admin" set.

also check that we used the embed-certs flag to embed client certificate data in the system:admin entry.

6. Create context for system:admin user into the new kubeconfig

Then, we need to create a context for our the system:admin.

A context element in a kubeconfig file is used to group access parameters under a convenient name. Each context has three parameters: cluster, namespace, and user. By default, the oc/kubectl command-line tool uses parameters from the current context to communicate with the cluster.

Based in the previous entry of the system:admin user, we can set up the context for our system:admin user:

$ oc config set-context system:admin --cluster=$(oc config view -o jsonpath='{.clusters[0].name}') --namespace=default --user=system:admin --kubeconfig=/tmp/$NEW_KUBECONFIG
Context "system:admin" modified.

7. Extraction of the Certificate Authority

Finally we need the CA of our Openshift cluster, because it is needed to complete the kubeconfig file for our system:admin.

We can extract this CA directly from our OpenShift cluster :

$ oc -n openshift-authentication rsh `oc get pods -n openshift-authentication -o name | head -1` \
cat /run/secrets/ > ingress-ca.crt

With the CA file we can add to the new kubeconfig with the following command:

$ oc config set-cluster $(oc config view -o jsonpath='{.clusters[0].name}') \
--server=$(oc config view -o jsonpath='{.clusters[0].cluster.server}') --certificate-authority=ingress-ca.crt --kubeconfig=/tmp/$NEW_KUBECONFIG --embed-certs
Cluster "api-cluster-xx-xx-sandbox-opentlc-com:6443" set.

and with that… voilà! We finished configuring our new kubeconfig file for our system:admin user!

Testing the new Kubeconfig file generated

8. Set current context to system:admin

Now that we have our kubeconfig, change the context to use our system:admin context generated in the previous step:

$ oc config use-context system:admin --kubeconfig=/tmp/$NEW_KUBECONFIG
Switched to context "system:admin".

9. Try the new kubeconfig file

In the last step we need to use the kubeconfig and test it in our OpenShift cluster.

Export the kubeconfig, log in with the system:admin user and do a privileged action like list nodes


$ oc login -u system:admin
Logged into "" as "system:admin" using existing credentials.

$ oc get nodes -o wide
NAME                        READY   STATUS    RESTARTS   AGE
console-56d5945c85-q7rjp    1/1     Running   0          2d3h
console-56d5945c85-x5669    1/1     Running   0          2d3h
downloads-9cb4cd587-dqdsr   1/1     Running   0          2d3h
downloads-9cb4cd587-m6pzz   1/1     Running   0          2d3h

NOTE: tested in OpenShift 4.4.17, but working in OCP3.x and OCP4.x

Automate the process!

I’m lazy sometimes to repeat the same process again and again, so I automated in a simple bash script and posted in a gist

Hope that helps!

Happy OpenShifting!