This section describe troubleshooting guidelines that should help you in troubleshooting your deployment and pipelines.
Pachyderm has a built-in logging system that collects information about events in your Pachyderm environment at pipeline, datum, and job level. See pachctl logs.
To troubleshoot the cluster itself, use the
kubectl tool troubleshooting tips. A few basic commands that you can use include the following:
- Get the list of all Kubernetes objects:
kubectl get all
- Get the information about a pod:
kubectl describe pod <podname>
The sections below provide troubleshooting steps for specific issues: