=== navidrome AFTER a removal that returned 200
created=2026-09-21T19:12:20.845623139Z started=2026-09-21T20:19:33.374056273Z restart=unless-stopped image=deluan/navidrome:0.64.0
compose project label: navidrome
app.yaml: ls: cannot access '/opt/docker/stacks/navidrome/app.yaml': No such file or directory
volume:   1 navidrome volume(s) left

TIMELINE, from the container's own metadata
-------------------------------------------
  21:12:05 local  POST /api/stacks/navidrome/remove {remove_hdd_data:true}  -> 409 (drive path
                  unresolvable, R-442's fail-closed guard — correct)
  21:12:11 local  POST /api/stacks/navidrome/remove {remove_hdd_data:false} -> 200,
                  volumes_removed = ['navidrome_navidrome_data']
  21:12:18 local  the harness's own leftovers check reported deployed=False and no containers
  21:12:20 local  A NAVIDROME CONTAINER WAS CREATED  (= container .Created, 19:12:20Z)
  22:19:33 local  ...and STARTED again by Docker's `restart: unless-stopped` at the B5 power cut
  22:27:20 local  the teardown found it: deployed=False, state=running

WHAT IS AND IS NOT ESTABLISHED
------------------------------
ESTABLISHED: the app's RECORD is gone (`app.yaml` absent, `deployed=false`), its VOLUME is gone,
and a container carrying `com.docker.compose.project=navidrome` was created two seconds after the
removal returned 200 and has run ever since. The controller probes it and calls it healthy
(`Health probe navidrome: API GET :4533/ping -> 200`).

NOT ESTABLISHED: WHAT created it. The controller was restarted several times later in the night
(knob changes and two power cuts) and its log no longer reaches 19:12:20Z. This is stated as an
observation, not a diagnosis — and it is the second time tonight that a restart destroyed the
evidence of the thing that mattered (see R-621).
