configurator / kubefs

Mount kubernetes metadata storage as a filesystem

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

kubefs

Build status

Mount kubernetes's metadata object store as a file system

Screenshot

Downloading

There are several options for downloading kubefs:

  • Download the latest release from GitHub

  • Build it directly from source using go get:

    go get github.com/configurator/kubefs

    Note: this is often broken because it not only gets the master branch of this repository - but also all the dependencies.

  • Build it directly from source by cloning this repository and running

    go build .

    to create a kubefs binary

  • Run it directly from source by cloning tihs repository and using

    go run . [<options>] <dir>

    instead of a kubefs binary.

Dependencies

For MacOS you will need osxfuse:

or you can install with brew formula:

$ brew cask install osxfuse

For Windows you will need WinFSP.

Usage

kubefs [<options>] <dir>

Mounts the default kubernetes cluster onto dir.

Options

  • -c, --kubeconfig:

    Like in kubectl, you can use the --kubeconfig flag to specify an alternate kube.config file, or pass the KUBECONFIG environment flag. All contexts in the passed config file will be available.

  • --show-json-files:

    Show json files in directory listings. By default, they are there but hidden.

  • --show-yaml-files=false:

    Hide yaml files in directory listings. If you don't combine this with --show-json-files, directories will all be empty. Files are still readable, just not listed in directory listings.

  • --pretty-json:

    When opening .json files, pretty-print them with newlines and indentation..

  • --readonly

    Mounts everything in read only mode so you'll feel safe (added in v0.4)

Files

Files in the mounted directory will have the following directory structure:

  • /<context>
    • all contexts available from kubeconfig should appear here, allowing you to browse multiple clusters without switching with kube config use-context.
  • /<context>/<object-type>
  • /<context>/<object-type>/<name>.yaml
    • for global objects that have no namespace, such as nodes,clusterroles or namespaces themselves
  • /<context>/<object-type>/<namespace>
  • /<context>/<object-type>/<namespace>/<name>.yaml
    • for namespaced objects such as pods, deployments, etc.

All object files end in .yaml. However, the file system has a secret - the extension can be changed to see the objects in different formats. Currently supported formats are:

  • .yaml
  • .json

These extra formats will not show up in directory listings, but are available to any application that tries to read them. Omitting the extension also works and returns the default format, yaml.

Ideas for exploring

  • Use find <dir> to see an entire listing of all kubernetes objects.

  • Open the directory in an IDE to look around

  • Snapshot the entire kubernetes object store by copying directory contents - though restoring isn't currently possible, you'd have a backup of each individual object.

Some more screenshots

Linux file listing

Browsing in IDE

Looking at yaml and json

Roadmap

  • List / get all kubernetes objects as files
  • Linux support
  • Mac support
  • Windows support
  • Delete objects
  • Edit existing objects
  • Create new objects
  • Support file system watchers, so IDEs know to reload the file after saving
  • Connect the file system watchers to kubernetes watchers so files can be reloaded when changed on the server
  • When writing files fails, add a comment to the top explaining the failure
  • When reading a missing file, allow the read with some dummy data as a result, which will help create new files in IDEs

About

Mount kubernetes metadata storage as a filesystem

License:MIT License


Languages

Language:Go 97.1%Language:Shell 2.9%