REPORT: v0.303.0 (R-897) delivered

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS

REPORT.md

@@ -1,4 +1,11 @@
-# REPORT — no code change (2026-10-07 day)
+# REPORT — controller v0.303.0: after a restart, a backup capture waits for the drive (2026-10-07, late evening)
 
-Only the shared rule file changed (`09` §3 decision 162: no hub image build or deploy in a session the operator does not
-attend; and rule 11, every helper prompt carries the brief's fences in full). Controller 0.302.0 stays on the three boxes.
+- R-897 (closed): for 10 minutes after the controller starts, a recovery-unit capture for an app on a drive runs only
+  once that drive is a live mount in the controller's own namespace (`/proc/self/mountinfo`); the 5-minute refresh
+  skips the app until then, a data run waits (5 s polls); after the window it runs and logs once at WARN. The agent's
+  bind order is unchanged. Measured cause: demo-hp 2026-10-07 15:16:28, `permission denied` in the second the agent
+  normalized the drive bind after a reboot.
+- `controller/internal/backup/driveready.go`; tests `TestDriveReady_*` (3; red-proved without the call).
+- MinAgent 0.131.0 (unchanged). Image `felhom-controller:0.303.0`; delivered by per-customer floors (demo-hp,
+  demo-felhom, tester-1; declared MinAgent; global floor untouched) — all three ran 0.303.0 at 18:50.
+- First live check: the kernel night 2026-10-07→08 (both demo boxes restart).