fix(reliability): converge runtime after profile restore (park dummy STA, re-raise monitors)

payload/user/remote_access/pager-webui/server.py

@@ -4894,6 +4894,21 @@ def h_profile_save(ctx):
     return (200, {'ok': True}) if ok else (502, {'ok': False})
 
 
+def _post_restore_converge():
+    """After any profile restore: re-apply invariants the restored config may
+    not carry. Factory-default configs enable dummy_radio0 (the phy0
+    channel-pinner) and may drop monitors; park and re-raise best-effort.
+    Uses ungated primitives — the restore is already inside its own gate."""
+    try:
+        _park_dummy_sta()
+    except Exception:
+        pass
+    try:
+        _raise_monitors()
+    except Exception:
+        pass
+
+
 def h_profile_restore(ctx):
     import mk8_events
     import mk8_gate
@@ -4911,6 +4926,7 @@ def h_profile_restore(ctx):
     ok = bool(result.get('ok'))
     if ok:
         device_run(['wifi', 'reload'], timeout=45)
+        _post_restore_converge()
     try:
         mk8_events.log_event('profile_restore',
                              sev='info' if ok else 'warn',
@@ -5001,6 +5017,16 @@ def _sta_uplink_enabled():
     return cfg.get('mode') == 'sta' and cfg.get('disabled') != '1'
 
 
+def _park_dummy_sta():
+    """Ungated primitive: disable dummy_radio0 + take wlan0 down. Callers
+    that already run inside a gate (post-restore convergence) use this to
+    avoid spawning a nested watchdog."""
+    device_run(['uci', 'set', 'wireless.dummy_radio0.disabled=1'])
+    device_run(['uci', 'commit', 'wireless'])
+    for iface in ('wlan0',):
+        rc, out, err = device_run(['ip', 'link', 'set', iface, 'down'], timeout=10)
+
+
 def _disable_sta_uplink():
     """Disable the dummy_radio0 STA without bouncing the radios. The UCI flag
     keeps it off across reboots/wifi reloads; taking wlan0 down immediately
@@ -5008,10 +5034,7 @@ def _disable_sta_uplink():
     tears down live APs and drops the monitors mid-assessment."""
     import mk8_gate
     mk8_gate.enter('uplink_disable')
-    device_run(['uci', 'set', 'wireless.dummy_radio0.disabled=1'])
-    device_run(['uci', 'commit', 'wireless'])
-    for iface in ('wlan0',):
-        rc, out, err = device_run(['ip', 'link', 'set', iface, 'down'], timeout=10)
+    _park_dummy_sta()
 
 
 def env_check():
@@ -7402,6 +7425,7 @@ if __name__ == '__main__':
         import mk8_profiles
         result = mk8_profiles.restore(name)
         device_run(['wifi', 'reload'], timeout=90)
+        _post_restore_converge()
         import mk8_events
         mk8_events.log_event('rollback', sev='warn',
                              msg='watchdog restored %s' % name,

tests/test_reliability_api.py

@@ -208,6 +208,17 @@ class ReliabilityApiTest(unittest.TestCase):
             self.assertEqual(calls['restored'], 'p1')
             reloads = [r for r in self.runs if r[0][:2] == ['wifi', 'reload']]
             self.assertEqual(len(reloads), 1)
+            # post-restore convergence parks the factory-enabled dummy STA
+            parked = [r for r in self.runs
+                      if r[0][:3] == ['uci', 'set',
+                                      'wireless.dummy_radio0.disabled=1']]
+            self.assertEqual(len(parked), 1)
+            downs = [r for r in self.runs if r[0][:4] == ['ip', 'link',
+                                                          'set', 'wlan0']]
+            self.assertTrue(downs)
+            raises = [r for r in self.runs if r[0][:4] == ['ip', 'link',
+                                                           'set', 'wlan0mon']]
+            self.assertTrue(raises)
             self.assertEqual(calls['events'][-1][0], 'profile_restore')
 
             status, result = server.h_profile_restore(