* refactor(supervisor): move backend setup behind the selected backend
Keep launch decoding, policy discovery and client construction with the
trusted backend while sharing supervisor admission and lifecycle handling.
Preserve VM policy identity validation and live credential handles.
Exercise alternate launch data through shared startup, readiness and
shutdown, including confirmation denial and resource cleanup.
Closes#4173
Signed-off-by: Shiju <shiju@nvidia.com>
* fix(supervisor): resolve backend setup lint failures
Box shared startup futures to keep callers' async state small.
Use the private module boundary for backend setup visibility and keep
test helpers and defaults explicit without changing their inputs.
Signed-off-by: Shiju <shiju@nvidia.com>
* refactor(supervisor): simplify selected backend startup
Signed-off-by: Shiju <shiju@nvidia.com>
* test(supervisor): keep imports before descriptor assertions
Signed-off-by: Shiju <shiju@nvidia.com>
* test(supervisor): preserve discovery failure startup coverage
Exercise backend discovery failure through shared startup before the
existing confirmation denial and successful shutdown cases. Assert that
startup preserves the error without launching a workload or readiness.
Signed-off-by: Shiju <shiju@nvidia.com>
---------
Signed-off-by: Shiju <shiju@nvidia.com>
John T. Myers