feat(reliability): risky-op preflight snapshots + detached rollback watchdog

This commit is contained in:
2026-08-22 13:56:25 -06:00
parent 7a27218149
commit a008bb9167
4 changed files with 182 additions and 0 deletions
+30
View File
@@ -0,0 +1,30 @@
#!/bin/sh
# Usage: mk8-watchdog.sh <profile> <interval> <fail_after> <healthy_after>
PROFILE="$1"; IV="${2:-5}"; FA="${3:-6}"; HA="${4:-6}"
DIR="/root/payloads/user/remote_access/pager-webui"
[ -f "$DIR/server.py" ] || DIR="/mmc/mk8/releases/current"
fails=0; oks=0; tripped=0
probe() {
curl -fsS -m 3 http://127.0.0.1:8080/ >/dev/null 2>&1 &&
{ ip link show wlan0mon >/dev/null 2>&1 ||
ip link show wlan1mon >/dev/null 2>&1; }
}
while true; do
if probe; then
fails=0
if [ "$tripped" = "1" ]; then
oks=$((oks + 1))
if [ "$oks" -ge "$HA" ]; then
/usr/bin/python3 "$DIR/server.py" --promote-snapshot "$PROFILE" >/dev/null 2>&1
exit 0
fi
fi
else
fails=$((fails + 1)); oks=0
if [ "$tripped" = "0" ] && [ "$fails" -ge "$FA" ]; then
tripped=1
/usr/bin/python3 "$DIR/server.py" --rollback-snapshot "$PROFILE" >/dev/null 2>&1
fi
fi
sleep "$IV"
done
@@ -0,0 +1,49 @@
"""Risky-operation gate: preflight config snapshot + detached rollback watchdog."""
import shlex
import subprocess
WATCHDOG = '/root/payloads/user/remote_access/pager-webui/mk8-watchdog.sh'
FAIL_AFTER = 6 # consecutive local-liveness failures -> rollback
HEALTHY_AFTER = 6 # consecutive successes after failure -> promote
INTERVAL = 5 # seconds between probes
# Dormant until an entrypoint (serve() / CLI ops) flips it on, so importing
# this module never snapshots or spawns anything.
ENABLED = False
def watchdog_decision(state):
"""state: {'fails': int, 'oks': int, 'tripped': bool,
'fail_after': 6, 'healthy_after': 6}
Returns (action, new_state): action in {'rollback','promote',None}."""
s = dict(state)
fa = s.get('fail_after', FAIL_AFTER)
ha = s.get('healthy_after', HEALTHY_AFTER)
if not s['tripped'] and s['fails'] >= fa:
return 'rollback', dict(s, tripped=True, oks=0)
if s['tripped'] and s['oks'] >= ha:
return 'promote', s
return None, s
def _spawn_watchdog(name):
cmd = ('setsid sh %s %s %d %d %d >/dev/null 2>&1 &'
% (shlex.quote(WATCHDOG), shlex.quote(name),
INTERVAL, FAIL_AFTER, HEALTHY_AFTER))
return subprocess.Popen(cmd, shell=True, start_new_session=True)
def enter(op):
"""Snapshot + spawn watchdog. Returns profile name or None when disabled."""
if not ENABLED:
return None
import mk8_profiles
name = mk8_profiles.auto_name(op)
mk8_profiles.snapshot(name)
_spawn_watchdog(name)
try:
import mk8_events
mk8_events.log_event('gate', msg='preflight snapshot %s' % name)
except Exception:
pass
return name
@@ -3526,6 +3526,8 @@ def _apply_radio1_ap(openap, wpa):
def h_pineap_wifi_set_ap(ctx): def h_pineap_wifi_set_ap(ctx):
import mk8_gate
mk8_gate.enter('ap_change')
body = ctx.body or {} body = ctx.body or {}
wpa = body.get('wpa') or {} wpa = body.get('wpa') or {}
openap = body.get('open') or {} openap = body.get('open') or {}
@@ -4350,6 +4352,8 @@ def _enterprise_boot_recover():
def h_attacks_deploy(ctx): def h_attacks_deploy(ctx):
import mk8_gate
mk8_gate.enter('attack_deploy')
body = ctx.body or {} body = ctx.body or {}
kind = (body.get('kind') or '').strip().lower() kind = (body.get('kind') or '').strip().lower()
if kind not in ('wpa', 'open', 'enterprise'): if kind not in ('wpa', 'open', 'enterprise'):
@@ -4439,6 +4443,8 @@ def _radio1_ap_active():
def h_attacks_stop(ctx): def h_attacks_stop(ctx):
import mk8_gate
mk8_gate.enter('attack_stop')
kind = ((ctx.body or {}).get('kind') or '') kind = ((ctx.body or {}).get('kind') or '')
stopped = [] stopped = []
if kind in ('wpa', 'open'): if kind in ('wpa', 'open'):
@@ -4890,6 +4896,8 @@ def _disable_sta_uplink():
keeps it off across reboots/wifi reloads; taking wlan0 down immediately keeps it off across reboots/wifi reloads; taking wlan0 down immediately
frees phy0's channel for wlan0mon. Never runs `wifi reload` here — that frees phy0's channel for wlan0mon. Never runs `wifi reload` here — that
tears down live APs and drops the monitors mid-assessment.""" 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', 'set', 'wireless.dummy_radio0.disabled=1'])
device_run(['uci', 'commit', 'wireless']) device_run(['uci', 'commit', 'wireless'])
for iface in ('wlan0',): for iface in ('wlan0',):
@@ -7189,6 +7197,8 @@ ROUTER.add('POST', r'/api/mode/release', h_mode_release)
def serve(): def serve():
import mk8_gate
mk8_gate.ENABLED = True
LIVE_STOP.clear() LIVE_STOP.clear()
HEALTH_STOP.clear() HEALTH_STOP.clear()
_recon_hopper_stop.set() _recon_hopper_stop.set()
@@ -7255,6 +7265,26 @@ if __name__ == '__main__':
except Exception as exc: # boot must never fail here except Exception as exc: # boot must never fail here
print(json.dumps({'error': str(exc)})) print(json.dumps({'error': str(exc)}))
sys.exit(0) sys.exit(0)
if '--rollback-snapshot' in sys.argv:
import mk8_gate
mk8_gate.ENABLED = True
name = sys.argv[sys.argv.index('--rollback-snapshot') + 1]
import mk8_profiles
result = mk8_profiles.restore(name)
device_run(['wifi', 'reload'], timeout=90)
import mk8_events
mk8_events.log_event('rollback', sev='warn',
msg='watchdog restored %s' % name,
meta=result)
print(json.dumps(result))
sys.exit(0)
if '--promote-snapshot' in sys.argv:
import mk8_gate
mk8_gate.ENABLED = True
name = sys.argv[sys.argv.index('--promote-snapshot') + 1]
import mk8_profiles
print(json.dumps({'promoted': mk8_profiles.promote_lastknown_good()}))
sys.exit(0)
signal.signal(signal.SIGTERM, _request_shutdown) signal.signal(signal.SIGTERM, _request_shutdown)
signal.signal(signal.SIGINT, _request_shutdown) signal.signal(signal.SIGINT, _request_shutdown)
serve() serve()
+73
View File
@@ -0,0 +1,73 @@
import os, sys, types, unittest
sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', 'payload', 'user', 'remote_access', 'pager-webui'))
import mk8_gate
class DecisionTest(unittest.TestCase):
def tick(self, state):
action, new = mk8_gate.watchdog_decision(state)
return action, new
def test_no_action_below_fail_threshold(self):
action, s = self.tick({'fails': 5, 'oks': 0, 'tripped': False})
self.assertIsNone(action)
self.assertFalse(s['tripped'])
def test_rollback_at_threshold(self):
action, s = self.tick({'fails': 6, 'oks': 0, 'tripped': False})
self.assertEqual(action, 'rollback')
self.assertTrue(s['tripped'])
self.assertEqual(s['oks'], 0)
def test_promote_after_recovery(self):
action, s = self.tick({'fails': 6, 'oks': 6, 'tripped': True})
self.assertEqual(action, 'promote')
def test_no_promote_before_recovery_threshold(self):
action, s = self.tick({'fails': 6, 'oks': 5, 'tripped': True})
self.assertIsNone(action)
class EnterTest(unittest.TestCase):
def setUp(self):
self.addCleanup(setattr, mk8_gate, 'ENABLED', mk8_gate.ENABLED)
self.spawned = []
def _install_fake_profiles(self):
snaps = []
fake = types.ModuleType('mk8_profiles')
fake.auto_name = lambda op: 'pre-%s-42' % op
fake.snapshot = lambda name: (snaps.append(name), True)[1]
old = sys.modules.get('mk8_profiles')
sys.modules['mk8_profiles'] = fake
self.addCleanup(sys.modules.__setitem__, 'mk8_profiles', old)
return snaps
def _capture_popen(self):
cmds = []
old = mk8_gate.subprocess.Popen
mk8_gate.subprocess.Popen = lambda cmd, **kw: cmds.append(cmd)
self.addCleanup(setattr, mk8_gate.subprocess, 'Popen', old)
return cmds
def test_enter_disabled_is_noop(self):
mk8_gate.ENABLED = False
cmds = self._capture_popen()
snaps = self._install_fake_profiles()
self.assertIsNone(mk8_gate.enter('ap_change'))
self.assertEqual(snaps, [])
self.assertEqual(cmds, [])
def test_enter_enabled_snapshots_and_spawns(self):
mk8_gate.ENABLED = True
cmds = self._capture_popen()
snaps = self._install_fake_profiles()
name = mk8_gate.enter('attack_deploy')
self.assertEqual(name, 'pre-attack_deploy-42')
self.assertEqual(snaps, [name])
self.assertEqual(len(cmds), 1)
self.assertIn('setsid sh', cmds[0])
self.assertIn(name, cmds[0])
if __name__ == '__main__':
unittest.main()