QA (public) » openQA Project (public) » openQA Infrastructure (public)

Observation¶

https://gitlab.suse.de/openqa/grafana-webhook-actions/-/jobs/283776 fails with

Running with gitlab-runner 13.5.0 (ece86343)
  on gitlab-worker1:leap15.1 HsXF8SXP
Preparing the "kubernetes" executor
00:00
Using Kubernetes namespace: gitlab
Using Kubernetes executor with image registry.opensuse.org/home/okurz/container/containers/tumbleweed:ipmitool-ping-nc ...
Preparing environment
00:03
Waiting for pod gitlab/runner-hsxf8sxp-project-4652-concurrent-0dshjz to be running, status is Pending
ERROR: Job failed (system failure): prepare environment: image pull failed: rpc error: code = Unknown desc = Error response from daemon: toomanyrequests: You have reached your pull rate limit. You may increase the limit by authenticating and upgrading: https://www.docker.com/increase-rate-limit. Check https://docs.gitlab.com/runner/shells/index.html#shell-profile-loading for more information

this is bad. We use registry.opensuse.org/home/okurz/container/containers/tumbleweed:ipmitool-ping-nc but looks like the gitlab CI runner itself might be using an image from docker hub?

Problem¶

We do not use docker hub and gitlab CI runners should not do as well – if they do.

Workaround¶

Run ARM worker recovery actions manually, e.g. ipmi-openqaworker-arm-2-ipmi power reset