A complete, runnable configuration that takes a freshly installed PCD Community Edition with one prepared host and builds a working region on it: a volume type and cluster blueprint with an NFS storage backend, host networking, a cluster, the hypervisor, image-library, and persistent-storage roles, then a provider network, a CirrOS image, a flavor, and an instance with a volume attached.
The walkthrough is the Community Edition guide on the Terraform Registry; the files here are the ones it renders.
source pcdctlrc # the RC file from Settings > API Access
export OS_INSECURE=true # Community Edition's self-signed certificate
cp terraform.tfvars.example terraform.tfvars
# edit terraform.tfvars
terraform init
terraform applyRole convergence takes several minutes per role; the whole apply is about twenty minutes on a small host.
terraform destroy removes everything, including the blueprint, but expect to
run it three times a few minutes apart: PCD refuses the storage role while the
deleted image's backing volume (image-<id>, in the service project) still
exists, so delete that volume with pcdctl volume delete between runs; and it
refuses the cluster and the host-configuration unassignment while a role's
deauthorization is still landing (HostClusterDeleteFailed,
HostInAuthState). The guide's Destroy section walks through it. The host
stays authorized and can be onboarded again.