argo doesn't work in openshift clusters

[nobody4t]

Summary

I deployed hello-world demo.
But I it got stuck for long in stage of pod creating.
I describe the pods and I got as below:

 Warning  FailedMount  48m (x189 over 17h)    kubelet  Unable to attach or mount volumes: unmounted volumes=[docker-sock], unattached volumes=[podmetadata docker-sock my-minio-cred default-token-jgg76]: timed out waiting for the condition

Diagnostics

What Kubernetes provider are you using?
OCP 4.6.19

What version of Argo Workflows are you running?
2.12.9

Paste a workflow that reproduces the bug, including status:
kubectl get wf -o yaml ${workflow} 

Just the hello-world demo in ocp

Paste the logs from the workflow controller:
kubectl logs -n argo $(kubectl get pods -l app=workflow-controller -n argo -o name) | grep ${workflow}

time="2021-03-04T09:12:44.061Z" level=info msg="Processing workflow" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:12:44.061Z" level=info msg="Updated phase  -> Running" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:12:44.062Z" level=info msg="Pod node hello-world-ldtts initialized Pending" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:12:44.121Z" level=info msg="Created pod: hello-world-ldtts (hello-world-ldtts)" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:12:44.136Z" level=info msg="Workflow update successful" namespace=argo phase=Running resourceVersion=8737051 workflow=hello-world-ldtts
time="2021-03-04T09:12:54.122Z" level=info msg="Processing workflow" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:12:54.135Z" level=info msg="Workflow update successful" namespace=argo phase=Running resourceVersion=8737209 workflow=hello-world-ldtts
time="2021-03-04T09:13:04.136Z" level=info msg="Processing workflow" namespace=argo workflow=hello-world-ldtts
time="2021-03-04T09:17:37.833Z" level=info msg="Performing periodic workflow GC"
time="2021-03-04T09:22:37.833Z" level=info msg="Performing periodic workflow GC"
time="2021-03-04T09:27:37.833Z" level=info msg="Performing periodic workflow GC"
time="2021-03-04T09:32:37.833Z" level=info msg="Performing periodic workflow GC"

---

Message from the maintainers:

Impacted by this bug? Give it a 👍. We prioritise the issues with the most 👍.

[nobody4t]

This issue shoule related to #826.
And I also tried the test here. It worked for me run a test pod with hostPath.
Any hints for this?

[alexec]

try using PNS executor

[nobody4t]

try using PNS executor

Sorry，I am new to this project.
Is there a link/doc for this?

[alexec]

https://github.com/argoproj/argo-workflows/blob/master/docs/workflow-executors.md

[nobody4t]

@alexec Thanks. I will try this.

[henrik-koren]

Deploy argo latest stable with the pns and k8sapi executor and encounter the same issues
Deploy argo versio v3.0.0-rc6 and the same issues reproduce

Deploy the argo application on an old cluster that have the docker engine running and it deploys successfully

Events:
  Type     Reason       Age                   From               Message
  ----     ------       ----                  ----               -------
  Normal   Scheduled    9m26s                 default-scheduler  Successfully assigned argocd/sparkly-bear to tkg01-workers-8s68h-75fcdc9cb6-kstrh
  Warning  FailedMount  7m23s                 kubelet            Unable to attach or mount volumes: unmounted volumes=[docker-sock], unattached volumes=[default-token-htwxr podmetadata docker-sock my-minio-cred]: timed out waiting for the condition
  Warning  FailedMount  2m51s                 kubelet            Unable to attach or mount volumes: unmounted volumes=[docker-sock], unattached volumes=[my-minio-cred default-token-htwxr podmetadata docker-sock]: timed out waiting for the condition
  Warning  FailedMount  72s (x12 over 9m26s)  kubelet            MountVolume.SetUp failed for volume "docker-sock" : hostPath type check failed: /var/run/docker.sock is not a socket file
  Warning  FailedMount  33s (x2 over 5m7s)    kubelet            Unable to attach or mount volumes: unmounted volumes=[docker-sock], unattached volumes=[podmetadata docker-sock my-minio-cred default-token-htwxr]: timed out waiting for the condition

This is from by ConfigMap
containerRuntimeExecutors:
----
- name: psn
  selector:
    matchLabels:
      workflows.argoproj.io/container-runtime-executor: psn

```

[nobody4t]

@alexec This works for me. Thanks.
@henrik-koren I think you only should setup containerRuntimeExecutor, containerRuntimeExecutors has a label which will only work for the matched scenario. This is my understand.

I will close this. Since my issue resolved.