Problems launching test in CP4DS cluster

[cwiering]

We seem to have a problem to launch the storage test right at the start when trying to login to OCP cluster. Manually logging into cluster works, ansible script has a problem. (supressed IDs and PWs here)

Manual login : [root@e1n1 ~]# oc login https://api.localcluster.fbond:6443/ -u kubeadmin -p ... --insecure-skip-tls-verify=true Login successful.

You have access to 117 projects, the list has been suppressed. You can list all projects with 'oc projects'

Using project "zen". [root@e1n1 ~]#

Ansible : [root@e1n1 ~]# run_k8s_storage_perf

PLAY [localhost] ***

TASK [ocp login using creds] *** fatal: [localhost]: FAILED! => {"changed": true, "cmd": "oc login https://api.localcluster.fbond:6443/ -u kubeadmin -p ... --insecure-skip-tls-verify=true", "delta": "0:00:00.099080", "end": "2023-01-10 17:45:23.952566", "msg": "non-zero return code", "rc": 1, "start": "2023-01-10 17:45:23.853486", "stderr": "error: dial tcp: lookup api.localcluster.fbond on x.x.x.x:53: server misbehaving - verify you have provided the correct host and port and that the server is currently running.", "stderr_lines": ["error: dial tcp: lookup api.localcluster.fbond on x.x.x.x:53: server misbehaving - verify you have provided the correct host and port and that the server is currently running."], "stdout": "", "stdout_lines": []}

PLAY RECAP ***** localhost : ok=0 changed=0 unreachable=0 failed=1 skipped=0 rescued=0 ignored=0

[root@e1n1 ~]#

We have a "disconnected environment" as Red Hat calls it. 8 node cluster, 3 master and 5 worker nodes. We have CP4DS 2.0.2.1 with CP4D 4.5.3 with ODF storage running on worker nodes.

[cwiering]

This is how it looked like when container was running (no port ?!) : [root@e1n1 ~]# podman ps -a CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES f3d93dd535c2 docker.io/library/registry:2.7 /etc/docker/regis... 7 weeks ago Up 7 weeks ago 0.0.0.0:5000->5000/tcp mirror-registry 02c79883b06c icr.io/cpopen/cpd/k8s-storage-perf:v1.0.0 17 hours ago Up 17 hours ago k8s-storage-perf

[bxu1999]

1. Is this run performed in an airgap OCP cluster?
2. Is the run kicked off from a pipeline job?
3. Can you follow the README file and manually run it to see if it works or not?

[cwiering]

1. "disconnected" = airgap, so YES
2. Because I'm not 100% sure, what a "pipeline job" is, I would say no :o|
3. Could you point me to the README file please ?

[bxu1999]

Hi @cwiering so this is a CPDS/Yosemite cluster, right? Those clusters have special OCP login method as far as I know, and we are not sure the existing supported OCP logins from this project would work. We have to test it first. Thanks.

[bxu1999]

Using the token method to perform OCP login to CPDS cluster should work, and I tested and it works fine as below.

• You just need to edit the params.yml file to provide the token on the line below:

  ocp_token: <required if user/password not available>

• Then run the Ansible Playbook:

  ......
  TASK [debug] ****************************************************************************************************************************************************
  ok: [localhost] => {
  "login_token.stdout_lines": [
      "Logged into \"https://api.localcluster.fbond:6443\" as \"system:serviceaccount:default:apadmin\" using the token provided.",
      "",
      "You have access to 94 projects, the list has been suppressed. You can list all projects with 'oc projects'",
      "",
      "Using project \"cpd-instance1\"."
  ]
  }
  ......