This document outlines the fields in pachyderm configs. This should act as a reference. If you wish to change a config value, you should do so via
If a field is not set, it will be omitted from JSON entirely. Following is an example of a simple config:
Following is a walk-through of all the fields.
A UUID giving a unique ID for this user for metrics.
Whether metrics is enabled.
v2.active_context specifies the name of the currently active pachyderm context, as specified in
Active Enterprise Context
v2.active_enterprise_context specifies the name of the currently active pachyderm enterprise context, as specified in
v2.contexts. If left blank the
v2.active_context value will be interpreted as the Active Enterprise Context.
A map of context names to their configurations. Pachyderm contexts are akin to kubernetes contexts (and in fact reference the kubernetes context that they're associated with.)
An integer that specifies where the config came from. This parameter is for internal use only and should not be modified.
host:port specification for connecting to pachd. If this is set, pachyderm will directly connect to the cluster, rather than resorting to kubernetes' port forwarding. If you can set this (because there's no firewall between you and the cluster), you should, as kubernetes' port forwarder is not designed to handle large amounts of data.
Trusted root certificates for the cluster, formatted as a base64-encoded PEM. This is only set when TLS is enabled.
A secret token identifying the current pachctl user within their pachyderm cluster. This is included in all RPCs sent by pachctl, and used to determine if pachctl actions are authorized. This is only set when auth is enabled.
The currently active transaction for batching together pachctl commands. This can be set or cleared via many of the
pachctl * transaction commands.
The name of the underlying Kubernetes cluster, extracted from the Kubernetes context.
The name of the underlying Kubernetes cluster's auth credentials, extracted from the Kubernetes context.
The underlying Kubernetes cluster's namespace, extracted from the Kubernetes context.
Cluster Deployment ID
The pachyderm cluster deployment ID that is used to ensure the operations run on the expected cluster.
Whether the context represents an enterprise server.
A mapping of
service name -> local port. This field is populated when you run explicit port forwarding (
pachctl port-forward), so that subsequent
pachctl operations know to use the explicit port forwarder.
This field is removed when the
pachctl port-forward operation completes. You might need to manually delete the field from your config if the process failed to remove the field automatically.
Last update: November 1, 2021