`ctl model serve`, its plan and model updates judged a file only against
the cluster's own llama.cpp, so a GPU node given a newer build could not
be served an architecture only that build knows: Inkling 2 Mini
(inkling2_loop) was refused for b10991 with cuda01 on b12606.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
engine_overrides give nodes picked by GPU class, id or labels their own
engine versions; node add --disable-engine keeps one off a node; what
vLLM cannot do on a node is a warning rather than a refused node; the
CUDA requirement follows the version; a hand-named binary is no longer
replaced by an installed one.
Refs #2 (items 4 and 7).
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>