ECP Vmware-Terraform

No longer maintained, replaced with Ezdemo

Overview

Setup HPE Ezmeral Container Platform (ECP) demo environment on Vmware vSphere with Terraform

This project makes it easy to setup HPE Container Platform demo/trial environments on Vmware, using AWS demo scripts to enable same learning/trial functionality.

Pre-requisities

Following tools are needed to use the scripts;

• Terraform download
• Packer download (required to create VM template)
• python3
• pip3
• ssh-keygen
• nc
• curl

Initialize with your Vmware credentials

vSphere environment (tested with ESXi 7.0.1, VCSA 7.0.1)

• Create
  • Datacenter
  • Cluster
  • Resource Pool
• DHCP should be enabled on the network and VMs should be accessible by your local machine
• DHCP should register DNS resolution (or manually provide DNS resolution for allocated IPs)

Update variables for your environment

Fill in the environment variables in ./etc/my_env.sh-template and save it as ./etc/my_env.sh

• centos_iso_url is a valid URL (ftp/http/https) to download CentOS ISO image (tested with 7.9)
• centos_iso_path is the pathname for local CentOS ISO image file (ie, "./CentOS-7-x86_64-Minimal-2009.iso")
• centos_iso_checksum is sha256 checksum for the file above (ie, "sha256:07b94e6b1a0b0260b94c83d6bb76b26bf7a310dc78d7a9c7432809fb9bc6194a")

Update ./etc/bluedata_infra.tfvars

• epic_dl_url (should be set in ./etc/my_env.sh)
• domain (should be set in ./etc/my_env.sh)
• (OPTIONAL)
  • Change project_id (also will be set to dns name for gateway)
  • Change options (create AD server, NFS server, External MapR cluster (not implemented yet), add GPU nodes (not implemented yet) etc)
  • Change VM sizes to fit in your environment
  • DO NOT enable RDP host, shouldn't be needed in local deployment (please open an issue if you think it is required)

Deploy

./bin/vmware_create_new.sh

This step might take somewhere between 45 minutes to 2 hours, please monitor the script output, and follow up guides in upstream repo.

Customizing resources and scenarios

You can edit ./etc/bluedata_infra.tfvars to choose some options, such as enabling RDP server, AD server, NFS server etc.

If you wish to re-configure, make changes and run terraform apply or re-run ./bin/vmware_create_new.sh to reflect these changes in the resources.

If you want clean installation and not all the demo scenarios, comment out following line in ./bin/vmware_create_new.sh:

mv "./etc/postcreate.sh_template" "./etc/postcreate.sh"

Alternatively, you can switch to "./hcp-demo-env-aws-terraform" folder and run scripts in ./bin directory (please run scripts on this upstream repo top level directory, ie, within ./hcp-demo-env-aws-terraform).

For example to install all available catalog images, you can run: ./bin/experimental/epic_catalog_image_install_all.sh

TODO

• Enable GPU nodes
• Enable MapR cluster creation

Please send comments/issues/suggestions through github (as we are not monitoring other places, such as Stackoverflow etc).