fix(recorders): kill stale standby container before on-demand respawn to prevent EADDRINUSE

services/mam-api/src/routes/recorders.js

@@ -801,6 +801,23 @@ router.post('/:id/start', requireRecorderEdit, async (req, res, next) => {
       }
     }
 
+    // If standby HTTP start failed and a stale container_id exists, kill it
+    // before spawning a new one — otherwise the new container gets EADDRINUSE
+    // because the old container is still holding the capture port.
+    if (!containerId && isStandby && recorder.container_id) {
+      console.log(`[recorders] killing stale standby container ${recorder.container_id} before respawn`);
+      try {
+        if (isRemote) {
+          await fetch(`${targetNodeApiUrl}/sidecar/${recorder.container_id}`, {
+            method: 'DELETE',
+            signal: AbortSignal.timeout(10000),
+          }).catch(() => {});
+        } else {
+          await dockerApi('DELETE', `/containers/${recorder.container_id}?force=true`).catch(() => {});
+        }
+      } catch (_) {}
+    }
+
     if (!containerId && isRemote) {
       // Remote node: delegate container lifecycle to that node's agent.
       const sidecarRes = await fetch(`${targetNodeApiUrl}/sidecar/start`, {