Guide to mmaictl (Command Line Interface)¶
GPU Cluster Manager offers and supports a command line interface package as part of it's product offering. This guide provides information on how to install, use, and troubleshoot the mmaictl command and it's subcommands. It also provides information on each subcommand, it's use, and related subcommands.
Installing mmaictl¶
To install mmai, download the package using:
Extract the mmaictl binary from the tar file using:
(Optional) Place the binary in a known location, such as /usr/local/bin:
Using mmaictl¶
To use mmaictl, from the directory where the binary exists, run ./mmaictl to list the options of the command:
./mmaictl
MMAI CLI, managing Kubernetes clusters with GPUs and AI workloads.
Find more information at: https://docs.memverge.com/AI
Usage:
mmaictl [command]
Resource Management
cluster Operations on clusters
department Operations on departments
node Operations on nodes
nodegroup Operations on nodegroups
project Operations on projects
Workload Management
compute Operations on compute assets
environment Operations on environment assets
job Operations on jobs
volume Operations on volumes
workload Operations on workloads
workspace Operations on workspaces
Additional Commands:
completion Generate the autocompletion script for the specified shell
context Operations for context
help Help about any command
kubectl Run kubectl commands
server Operations for MMAI servers
Flags:
-c, --config string Path to mmaictl config directory (default "/root/.mmaictl")
-h, --help help for mmaictl
-v, --version version for mmaictl
Use "mmaictl [command] --help" for more information about a command.
Troubleshooting mmaictl¶
You may encounter the following when using mmaictl. The message is displayed because the cluster admin must first login to the cluster to perform actions. See mmaictl server login for details.