RHEL for edge

Build ISO for edge devices

• prepare host with RHEL8.5+ - 2 CPUs, 4GB memory and 20GB of storage. It may be a VM, CNV VM or physical appliance. Web server running must be reachable by edge devices.
• The root folder should have more than 15GB available, you can extend it by:
  • resize the virtual machine disk before that step make sure the VM isn't running

    qemu-img resize <img> +15G
  

  • run the VM and resize the root partition and the filesystem (inside the VM):

    growpart /dev/sda 3
    resize2fs /dev/sda3
  

• use a user with sudo - make sure user does not need to enter password each time (set NOPASSWD in sudoers file)
• make sure selinux is running in permissive mode
• register the system

  sudo subscription-manager register --username <redhat_login_username> --password <redhat_login_password> --auto-attach
  

• execute flotta-agent-sti.sh for creating rhel4edge image for edge-device

  ./flotta-agent-sti.sh -a <image-server-address>  -i <image-name> -o <operator_host_name:operator_port> -v
  

Upgrade device

In the builder machine:

• execute flotta-agent-sti-upgrade.sh for creating rhel4edge commit on top of image or commit for upgrading edge-device

  ./flotta-agent-sti-upgrade.sh -a <server_adress> -i <image_name> -o <operator_host_name:operator_port> -p <parent_commit> -v
  
  

In the edge device machine:

• update the remote url of edge (you might see it under /etc/ostree/remotes.d/edge.conf) to point the url of the new commit

• for checking there is an updated image available

  rpm-ostree update --preview
  

• configure automate rolling upgrade by using greenboot

• copy the script greenboot-health-check.sh to /etc/greenboot/check/required.d/greenboot-health-check.sh An health check script that must not fail (if they do, GreenBoot will initiate a rollback)

  chmod +x /etc/greenboot/check/required.d/greenboot-health-check.sh
  

• copy the script bootfail.sh to /etc/greenboot/red.d/bootfail.sh scripts that should be run after GreenBoot has declared the boot as failed.

  chmod +x /etc/greenboot/red.d/bootfail.sh
  

• for updating run

  rpm-ostree update
  rpm-ostree status 
  systemctl reboot
  

  if the greenboot would be failed after several runs greenboot would roll back automatically

• in order to see greenboot logs:

  systemctl status greenboot-status