The deployment guide being written now is deliberately a step-by-step lab walkthrough: concrete addresses, concrete commands, followable end to end so a manual lab can be built from it and the setup is testable. That is the right first target.
It is not the same document as an operator-facing reference: one that states requirements and options without prescribing a single topology, so a user can map them onto their own network. That layer is deferred on purpose — write it once the walkthrough has proven which parts are essential and which were lab-specific.
What the reference layer should eventually cover, phrased as requirements rather than a recipe:
- The scenario matrix — existing DHCP on the same L2 (proxyDHCP coexistence), existing DHCP reached by relay, isolated segment with bootd serving leases — and how to tell which one a site is.
- Network prerequisites the user owns: the provisioning L2 per site, BMC reachability, cluster-to-site L3 baseline.
- The tracker/seeder connectivity matrix as requirements: targets must reach the announce URL and the seeder BT endpoint, and the address the tracker records must be real (no NAT). Both fulfilments are valid — a routed L3 path to a stable cluster Service address, or attaching tracker/seeder to the target subnet — without declaring one canonical.
- The full port table (67/4011 DHCP, 69 TFTP, boot config, agent API, 6969 tracker, 16881 BT) and what must reach what.
- The image contract (bootable images ship their own removable-media bootloader) and the Secure Boot chain.
- Sizing/scale guidance once real numbers exist.
Written in ASD-STE100 English, same as the rest of docs/.
The deployment guide being written now is deliberately a step-by-step lab walkthrough: concrete addresses, concrete commands, followable end to end so a manual lab can be built from it and the setup is testable. That is the right first target.
It is not the same document as an operator-facing reference: one that states requirements and options without prescribing a single topology, so a user can map them onto their own network. That layer is deferred on purpose — write it once the walkthrough has proven which parts are essential and which were lab-specific.
What the reference layer should eventually cover, phrased as requirements rather than a recipe:
Written in ASD-STE100 English, same as the rest of
docs/.