VMs are provisioned via Terraform, and within that VM the clusters are setup using Ansible and other scripts.
GitHub rate limits mean repository creation may fail during bulk VM creation. If that happens, the VM creation will continue, and we can provision the repositories separately.
- Login to the workshop VM
- Navigate to the common workshop admin directory, and verify you have the
latest scripts.
sudo su - workshop cd one-summit-22-workshop git pull --ff-only
- Adjust the number of VMs to create by editing
infra/compute_instances.tf
, setting thenum_vms
parameter (more info), then execute terraform:cd infra terraform plan # you should see that new VMs will be created terraform apply
- The result will give you the name and IP of the new VMs. They are also
available via
gcloud compute instances list
. - For any new VMs, you may need to create the GitHub repositories via
./scripts/repos.sh create $PARTICIPANT
- The
repos.sh
script and GitHub token are available on the VM; we could run./repos.sh create $(hostname)
to automatically create the repositories. If we do this, the script may fail due to repostories already existing, or due to GitHub rate limits. So, we need to be sure it failing does not abort the rest of setup. We also need to be prepared to run the script again from the workshop VM to provision any repositories that failed to be created.