- mac: cnvrgctl-darwin-x86_64
curl -#o /usr/local/bin/cnvrgctl \ https://cnvrg-public-images.s3-us-west-2.amazonaws.com/cnvrgctl-darwin-x86_64 \ && chmod +x /usr/local/bin/cnvrgctl \ && cnvrgctl completion bash > /usr/local/etc/bash_completion.d/cnvrgctl
- linux: cnvrgctl-linux-x86_64
curl -#o /usr/local/bin/cnvrgctl \ https://cnvrg-public-images.s3-us-west-2.amazonaws.com/cnvrgctl-linux-x86_64 \ && chmod +x /usr/local/bin/cnvrgctl \ && cnvrgctl completion bash > /etc/bash_completion.d/cnvrgctl
Prerequisite
- VM or Bare metal Ubuntu 20.04 server with 32 CPUs, 64GB memory, 500GB storage
- In case of VM, bridged network (preferred) or nat network between VM and the host
- either root user or regular user with sudo access
- SSH access to the server either by ssh key or password
Deploy single node K8s cluster for cnvrg deployment
# access the server with ssh password
cnvrgctl cluster up --host=<SERVER-IP> --ssh-user=<SSH-USER> --ssh-pass=<SSH-PASS>
# access the server with ssh key
cnvrgctl cluster up --host=<SERVER-IP> --ssh-user=<SSH-USER> --ssh-pass=</path/to/private/key>
Cleanup
# access the server with ssh password
cnvrgctl cluster destroy --host=<SERVER-IP> --ssh-user=<SSH-USER> --ssh-pass=<SSH-PASS>
# access the server with ssh key
cnvrgctl cluster destroy --host=<SERVER-IP> --ssh-user=<SSH-USER> --ssh-pass=</path/to/private/key>
There are two ways for importing cnvrg stack images into your internal air gap docker registry
- Using
cnvrgctl
- Using bash scripts generated by
cnvrgctl
To see all the available options, run cnvrgctl -h
, for example, for listing all the required images for whitening, run cnvrgctl images dump --list
- Pull all the images on the linux machine connected to internet
# the cnvrg pull user/pass should be provided as a part of your license subscription cnvrgctl images pull --registry-user=<CNVRG-PULL-USER> --registry-pass=<CNVRG-PULL-KEY>
- Save all images to local disk as an images archives
cnvrgctl images save
- Run whitening process, after that copy all the images archives to internal linux machine, once copied, run
cnvrgctl images load
- Tag images with your internal docker registry and docker repo
cnvrgctl images tag \ --registry=<INTERNAL-PRIVATE-REGISTY> \ --registry-repo=<INTERNAL-REPO-FOR-CNVRG-IMAGES>
- Push tagged images into your internal docker registry
cnvrgctl images push \ --registry=<INTERNAL-PRIVATE-REGISTY> \ --registry-repo=<INTERNAL-REPO-FOR-CNVRG-IMAGES> \ --registry-user=<INTERNAL-REGISTY-USER> \ --registry-pass=<INTERNAL-REGISTY-PASSWORD>
- Generate and execute bash script for pulling all the images on the linux machine connected to internet
cnvrgctl images dump --pull
- Generate and execute bash script for saving all images to local disk as an images archives
cnvrgctl images dump --save
- Run whitening process, after that, copy all the images archives to internal linux machine, once copied, generate bash script for image load
cnvrgctl images dump --load
- Generate tag script and run it to load images into yours internal registry
cnvrgctl images dump --tag \ --registry=<INTERNAL-PRIVATE-REGISTY> \ --registry-repo=<INTERNAL-REPO-FOR-CNVRG-IMAGES>
- Generate push script to push tagged images into your internal docker registry
cnvrgctl images dump --push \ --registry=<INTERNAL-PRIVATE-REGISTY> \ --registry-repo=<INTERNAL-REPO-FOR-CNVRG-IMAGES>