Skip to content

Latest commit

 

History

History
123 lines (87 loc) · 5.64 KB

nuctl.md

File metadata and controls

123 lines (87 loc) · 5.64 KB
Raw

Nuclio CLI

In this document

Overview

nuctl is Nuclio's command-line interface (CLI), which provides command-line access to Nuclio features.

Download and Installation

To install nuctl, simply visit the Nuclio releases page and download the appropriate CLI binary for your environment (for example, nuctl-<version>-darwin-amd64 for a machine running macOS).

You can use the following command to download the latest nuctl release:

curl -s https://api.github.com/repos/nuclio/nuclio/releases/latest \
			| grep -i "browser_download_url.*nuctl.*$(uname)" \
			| cut -d : -f 2,3 \
			| tr -d \" \
			| wget -O nuctl -qi - && chmod +x nuctl

Usage

After you download nuctl, use the --help option to see full usage instructions:

nuctl --help

The running platform

nuctl automatically identifies the platform from which it's being run. You can also use the --platform CLI flag to ensure that you're running against a specific platform, as explained in this reference:

Local Docker

To force nuctl to run locally, using a Docker daemon, add --platform local to the CLI command.

For an example of function deployment using nuctl against Docker, see the Nuclio Docker getting-started guide.

Kubernetes

To force nuctl to run against a Kubernetes instance of Nuclio, add --platform kube to the CLI command.

When running on Kubernetes, nuctl requires a running registry on your Kubernetes cluster and access to a kubeconfig file.

For an example of function deployment using nuctl against a Kubernetes cluster, see the Nuclio Kubernetes getting-started guide.

Invoking a function using nuctl on Kubernetes

While in most cases using nuctl on the Local Docker platform enables you to invoke your function by running the nuctl invoke command on the host, in Kubernetes platform this is a bit trickier - and will most likely not work by default for some configurations. This means, for example, that nuctl invoke command will not be able to work unless your function is explicitly exposed or reachable from wherever you run nuctl invoke.

See exposing a function for more details.

For your convenience, when deploying a function using nuctl, exposing it via a NodePort can be easily done by using the CLI arguments --http-trigger-service-type=nodePort.

Redeploying functions

Redeploy allows deploying functions that have pre-built images, and have been deployed in the past. This is useful in cases when you change your platform (e.g. Kubernetes) environment, for backup and restore purposes, and so on. In case a full deployment is needed, along with rebuilding the function images, use nuctl deploy command.

Currently redeploy command is the only command that uses dashboard API. Namely, it uses a Patch request.

Use-cases:

  • redeploy imported functions (for instance, after platform migration, backup and restore, etc.)
  • simply redeploy an already existing function

Functions can be successfully redeployed to either ready (default) or scaledToZero states. To be able to redeploy function as scaledToZero, minimum replicas of function should be equal to zero.

Command flags:

  • verify-external-registry - functions will be redeployed only if registry is external
  • save-report - allows to save redeployment report to a file. Path can be specified via report-file-path argument
  • report-file-path - report file path. Default is ./nuctl-redeployment-report.json
  • from-report - redeploy failed retryable functions from report
  • imported-only - redeploy only imported functions
  • desired-state - redeploy to specific status. Choose any of those statuses: [scaledToZero, ready].
  • exclude-projects - exclude projects
  • exclude-functions - exclude functions
  • exclude-functions-with-gpu - exclude only functions with GPU
  • wait - waits for redeployment to be completed
  • wait-timeout - waits timeout duration for the function redeployment

Redeployment report interpretation

Redeployment report contains 3 lists with function names - Success, Skipped, Failed.

  • Success: a list containing the names of functions that were successfully redeployed.
  • Skipped: a list containing function names that the redeployment process has skipped.
  • Failed: a list containing the functions that failed redeploying. For each failed function, the report contains a failure descriptions and a boolean variable that shows whether rerunning the command can help.

If retryable == false in your report, there are possible reasons:

  • No image field found in function config spec (check spec.image)
  • Desired state is scaledToZero but minReplicas > 0
  • imported-flag was used in redeploy command and function not in imported state

Redeploying imported functions

To redeploy only imported functions use --imported-only flag.

An imported function can be redeployed to the state it had before being imported. To do so, the function should be exported with its previous state, meaning the function has a nuclio.io/previous-state annotation.