Manual Installation of CML on instance using the ubuntu/images/hvm-ssd/ubuntu-focal-20.04-amd64-server-20230918

[osacalsaw]

We are able to deploy a bare metal instance on the console using the ubuntu/images/hvm-ssd/ubuntu-focal-20.04-amd64-server-20230918 as per the CML terraform code. But we are not able to deploy using Terraform as per the guidelines (The Terraform user is prevented for creating resources due to a Service Control Policy that due to organizational policies, cannot be removed). With the bare metal instance deployed on the console, is there a way to manually deploy and install the CML software and refplat files on the instance? Are there any directions for this? We have the refplat files and CML software in an S3 bucket.

[rschmied]

very roughly, and from the top of my head (as this is venturing into uncharted/unsupported land):

• copy the cml2*.pkg file to the server
• extract the .deb packages from the cml2*.pkg using tar
• install the cml*.deb, patty*.deb and iol-tools*.deb packages on the server
• look at the cml.sh file from this repository to get some ideas how the configuration works
  • modules/deploy/data/cml.sh
  • modules/deploy/data/virl2-base-config.yml
• the initial configuration services consumes the data from the virl2-base-config.yml file
• you need to copy all the node and image definitions and actual images from S3 to the VM (suggest to use aws s3 sync), they go into /var/lib/libvirt/images/virl2-base-images and /var/lib/libvirt/images/node-definitions.
• licensing of the system is best done then via the actual UI

it's also up to you at this point how to provide IP access to the system...