Version
redhat-actions/oc-login@v1
Describe the bug
We experienced a race condition where we had two separate pipelines run at the same time on the same runner, changing the KUBECONFIG's current-context:
app1 logs in
Fri, 05 Sep 2025 16:04:20 GMT Run redhat-actions/oc-login@v1
Fri, 05 Sep 2025 16:04:20 GMT /home/github/runner01/_work/_temp/openshift-bin/oc login
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner01/_work/_temp/openshift-bin/oc config current-context
Fri, 05 Sep 2025 16:04:21 GMT app1/***/system:serviceaccount:app1:app1-sa
Fri, 05 Sep 2025 16:04:21 GMT Set current context's namespace to "app1"
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner01/_work/_temp/openshift-bin/oc config set-context app1/***/system:serviceaccount:app1:app1-sa --namespace=app1
Fri, 05 Sep 2025 16:04:21 GMT Context "app1/***/system:serviceaccount:app1:app1-sa" modified.
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner01/_work/_temp/openshift-bin/oc config view --flatten
Fri, 05 Sep 2025 16:04:21 GMT *** Suppressing command output
Fri, 05 Sep 2025 16:04:21 GMT Writing out Kubeconfig to /home/github/runner01/_work/app1/app1/kubeconfig.yaml
Fri, 05 Sep 2025 16:04:21 GMT Kubeconfig contents
Fri, 05 Sep 2025 16:04:21 GMT Exporting KUBECONFIG=/home/github/runner01/_work/app1/app1/kubeconfig.yaml
At the same time, another app (app2) logs in (on a different runner though on the same VM)
Fri, 05 Sep 2025 16:04:20 GMT Run redhat-actions/oc-login@v1
Fri, 05 Sep 2025 16:04:20 GMT /home/github/runner02/_work/_temp/openshift-bin/oc login
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner02/_work/_temp/openshift-bin/oc config current-context
Fri, 05 Sep 2025 16:04:21 GMT app2/***/system:serviceaccount:app2:app2-sa
Fri, 05 Sep 2025 16:04:21 GMT Set current context's namespace to "app2"
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner02/_work/_temp/openshift-bin/oc config set-context app2/***/system:serviceaccount:app2:app2-sa --namespace=app2
Fri, 05 Sep 2025 16:04:21 GMT Context "app2/***/system:serviceaccount:app2:app2-sa" modified.
Fri, 05 Sep 2025 16:04:21 GMT /home/github/runner02/_work/_temp/openshift-bin/oc config view --flatten
Fri, 05 Sep 2025 16:04:21 GMT *** Suppressing command output
Fri, 05 Sep 2025 16:04:21 GMT Writing out Kubeconfig to /home/github/runner02/_work/app2/app2/kubeconfig.yaml
Fri, 05 Sep 2025 16:04:21 GMT Kubeconfig contents
Fri, 05 Sep 2025 16:04:21 GMT Exporting KUBECONFIG=/home/github/runner02/_work/app2/app2/kubeconfig.yaml
app1 has it's context changed and is now in another app2 namespace, which fails to deploy
Fri, 05 Sep 2025 16:04:21 GMT Run oc project
Fri, 05 Sep 2025 16:04:21 GMT oc project
Fri, 05 Sep 2025 16:04:21 GMT shell: /usr/bin/bash -e {0}
Fri, 05 Sep 2025 16:04:21 GMT env:
Fri, 05 Sep 2025 16:04:21 GMT NAMESPACE: app1
Fri, 05 Sep 2025 16:04:21 GMT IMAGE_NAME: app1
Fri, 05 Sep 2025 16:04:21 GMT KUBECONFIG: /home/github/runner01/_work/app1/app1/kubeconfig.yaml
Fri, 05 Sep 2025 16:04:21 GMT Using project "app2" on server "https://api.<CLUSTER>:6443".
...
Fri, 05 Sep 2025 16:04:21 GMT Run oc apply...
Fri, 05 Sep 2025 16:04:21 GMT from server for: "STDIN": configmaps "app1" is forbidden: User "system:serviceaccount:app2:app2-sa" cannot get resource "configmaps" in API group "" in the namespace "app1"
...
Steps to reproduce, workflow links, screenshots
See above, just using the oc-login action at the same time for two or more Jobs, creating a race condition.
Version
redhat-actions/oc-login@v1Describe the bug
We experienced a race condition where we had two separate pipelines run at the same time on the same runner, changing the KUBECONFIG's
current-context:app1logs inAt the same time, another app (
app2) logs in (on a different runner though on the same VM)app1has it's context changed and is now in anotherapp2namespace, which fails to deploySteps to reproduce, workflow links, screenshots
See above, just using the
oc-loginaction at the same time for two or more Jobs, creating a race condition.