cmd/

Table of Contents

• Executables
• Installing from GitHub
• Installing from source

Executables

NOTE: installed CLI executable is named ais.

Installing from GitHub

Generally, AIStore (cluster) requires at least some sort of deployment process or sequence.

Standalone binaries, on the other hand, can be built from source or installed directly from the latest or previous GitHub releases.

NOTE: binary installation is supported only for the linux-amd64 platform.

In particular:

$ ./scripts/install_from_binaries.sh --help

NAME:
  install_from_binaries.sh - install 'ais' (CLI) and 'aisloader' from release binaries

USAGE:
  ./install_from_binaries.sh [options...]

OPTIONS:
  --tmpdir <dir>        work directory, e.g. /root/tmp
  --dstdir <dir>        installation destination
  --release             e.g., v1.3.15, v1.3.16, latest (default: latest)
  --completions         install and enable _only_ CLI autocompletions (ie., skip installing binaries)

NOTE: For CLI, the script will also enable auto-completions. CLI can be used without (bash, zsh) auto-completions but, generally, using auto-completions is strongly recommended.

Example: download 'ais' and 'aisloader' binaries from the latest release

$ ./install_from_binaries.sh --dstdir /tmp/qqq

Upon execution, the two specific linux-amd64 binaries, ready for usage, will be placed in /tmp/qqq destination.

Installing from source

CLI

The preferable way is to use Makefile:

$ make cli

builds AIS CLI from the local aistore repository and installs it in your $GOPATH/bin.

NOTE: installed CLI binary is named ais.

Alternatively, you could also use go install:

$ go install github.com/NVIDIA/aistore/cmd/cli@latest` && mv $GOPATH/bin/cli $GOPATH/bin/ais

To install CLI auto-completions, you could also, and separately, use cmd/cli/install_autocompletions.sh

aisloader

Makefile way:

$ make aisloader

But again, you could also use go install:

$ go install github.com/NVIDIA/aistore/cmd/aisloader@latest

xmeta

xmeta is a low-level utility to format (or extract and show) assorted AIS control structures - see usage.

For command line options and usage examples, simply run xmeta with no arguments:

$ xmeta
Usage of xmeta:
  -f string ...
...
Examples:
        # Smap:
        xmeta -x -in=~/.ais0/.ais.smap      - extract Smap to STDOUT
...

To install, run:

$ make xmeta

OR, same:

$ cd cmd/xmeta
$ go install

ishard

ishard is a AIS integrated utility to create well-formed shards from the original dataset - see usage.

For command line options and usage examples, simply run ishard with no arguments:

$ ishard
Usage of ishard:
  ...
...

To install, run:

$ make ishard

You could also use go install:

$ go install github.com/NVIDIA/aistore/cmd/ishard@latest