Keep OPNsense telemetry when the script worker cannot start and build it for FreeBSD 14.1

A worker linked on the FreeBSD 15 test guest needs FBSD_1.9 and does not load on
OPNsense 24.7; the agent then retried the job session every 30 s and never sent a
heartbeat or inventory. The agent now pauses script capabilities for an hour and
reconnects without them. tools/dev_opnsense_worker.py links the worker against the
official FreeBSD 14.1 pkgbase libraries. The installer checks that the worker loads
before binding it and reports success only after a heartbeat.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-11-21 12:17:34 +01:00
co-authored by Claude Opus 5.5
parent 19163a2efe
commit 93429e4394
8 changed files with 216 additions and 15 deletions
+14 -4
View File
@@ -20,9 +20,13 @@ STATE = Path('/var/db/discovery-agent')
CA = '/usr/local/etc/discovery-agent/ca.pem'
HOST = 'discovery.ease-technologies.de'
MEDIA = 'application/x-protobuf'
VERSION = '0.1.0-opnsense.6'
VERSION = '0.1.0-opnsense.7'
WORKER = Path('/usr/local/libexec/discovery-agent/rmm-opnsense-jobs')
BINDING = Path('/usr/local/etc/discovery-agent/job-binding.json')
# A worker that cannot start (e.g. built for a newer FreeBSD) must not stop
# heartbeat and inventory. Script capabilities pause until this moment.
WORKER_RETRY = 3600
worker_paused_until = 0.0
def post(ctx, path, data, media):
@@ -177,11 +181,12 @@ def session(db, identity, boot, instance):
def run_channel(db, identity, boot, instance, stack):
global worker_paused_until
ctx = context()
ctx.load_cert_chain(STATE/'device.pem', STATE/'device.key')
generation = 0
jobs = None
enabled = BINDING.is_file() and WORKER.is_file()
enabled = BINDING.is_file() and WORKER.is_file() and time.monotonic() >= worker_paused_until
def send(kind, payload, captured=None):
# A lost inventory ACK replays the exact durable envelope, including its
@@ -249,8 +254,13 @@ def run_channel(db, identity, boot, instance, stack):
p.stdout.close()
stack.callback(close)
jobs = Jobs(p)
if jobs.read()!=b'ready':
raise ValueError('job worker unavailable')
try:
ready = jobs.read()==b'ready'
except ValueError:
ready = False
if not ready:
worker_paused_until = time.monotonic()+WORKER_RETRY
raise ValueError('job worker unavailable; scripts paused, telemetry continues')
last = None
previous = None
check_cert = time.monotonic()+12*3600