Commit Graph
13 Commits
Author SHA1 Message Date
Christian Manivong a65a3ed6e3 refactor(config): take the revision contract from napalm-device-types
driver / standard (pull_request) Successful in 17s
driver / lint (pull_request) Successful in 15s
driver / types (pull_request) Successful in 1m0s
driver / test (3.12) (pull_request) Successful in 1m11s
driver / test (3.13) (pull_request) Successful in 1m3s
driver / lowest (pull_request) Successful in 35s
driver / build (pull_request) Successful in 29s
driver / security (pull_request) Successful in 1m1s
CI / driver (pull_request) Successful in 2m44s
driver / publish (pull_request) Skipped
napalm-device-types 4.38.0 declares ConfigRevisionsMixin: the listing,
the reading, ConfigRevisionDict and the exceptions, and the revert as a
template. The driver keeps the device communication: the three backup
API calls, the revert step, and config_revision_content(), which leaves
out the <revision> block every save writes into config.xml and the
indentation between elements.

The template reads the newest revision after the revert and checks that
it holds the revision's configuration. OPNsense answers "reverted" even
when restoreBackup() could not load the file and kept the previous one;
that now raises ConfigRevertFailed instead of passing. A refusal of the
firewall raises ConfigRevertFailed too. An id the history does not list
still never reaches a request path.

Refs NetOrk/netork#2511
2026-11-13 22:33:58 +01:00
Christian Manivong 5a20023c54 feat(config): list, read and revert the firewall's config revisions
driver / lint (pull_request) Successful in 14s
driver / standard (pull_request) Successful in 17s
driver / types (pull_request) Successful in 57s
driver / test (3.13) (pull_request) Successful in 1m4s
driver / test (3.12) (pull_request) Successful in 1m10s
driver / lowest (pull_request) Successful in 30s
driver / build (pull_request) Successful in 29s
driver / security (pull_request) Successful in 1m4s
CI / driver (pull_request) Successful in 2m25s
driver / publish (pull_request) Skipped
OPNsense has no API that takes an uploaded config.xml: the backup API
lists the history in /conf/backup, hands out a revision and reverts to
one that is already on the firewall. A full replace exists only on the
legacy page diag_backup.php. So an application restores a configuration
by reverting to the revision that holds it (NetOrk/netork#2511).

get_config_revisions() lists the history newest first: id, time in UTC
(time_iso is in the firewall's zone), description, user and size.
get_config_revision(id) returns one revision's config.xml.
revert_config_revision(id) calls revert_backup and refuses an id that is
not in the history now; an id that is no file name of the history never
reaches a request path. A revert only saves config.xml as a new newest
revision: no service is reloaded and the firewall does not reboot.

load_replace_candidate is None instead of NAPALM's stub that raised, so
a capability check sees that the driver replaces nothing. The helper
that finds the newest backup for rollback() moves to the new module,
which keeps opnsense.py under its recorded size.

Refs NetOrk/netork#2511
2026-11-13 22:21:58 +01:00
Christian Manivong 6fdc3ff2c7 feat(time-sync): read the time servers and zone of config.xml (NetOrk/netork#2363)
driver / standard (pull_request) Successful in 17s
driver / lint (pull_request) Successful in 14s
driver / types (pull_request) Successful in 1m1s
driver / test (3.12) (pull_request) Successful in 1m9s
driver / test (3.13) (pull_request) Successful in 1m3s
driver / lowest (pull_request) Successful in 36s
driver / build (pull_request) Successful in 27s
driver / security (pull_request) Successful in 58s
CI / driver (pull_request) Successful in 2m33s
driver / publish (pull_request) Skipped
netOrk aligns every device with one list of time servers and a zone per
site (NetOrk/netork#2360) through napalm-device-types 4.35.0's
TimeSyncMixin. OPNsense keeps both in config.xml, <system><timeservers>
and <system><timezone>, set on the legacy pages services_ntpd.php and
system_general.php. No API controller reads or writes them, on 26.7 and
on master, so get_time_sync() reads them from the configuration download
get_config already uses and asks /api/ntpd/service/status whether ntpd
follows a system peer.

The driver writes neither value: time_sync_writes is empty and
configure_time_sync is None. Restoring an edited config.xml or running
the initial setup wizard would rewrite far more than the clock, on a box
that is often the gateway and its LAN's time server. TIME_SERVERS_MENU
and TIME_ZONE_MENU say where a person sets them, as SSH_SETTINGS_MENU
does for the SSH algorithms.

get_ntp_servers() now returns the same configured list. It read a
"peers" field the status endpoint does not have (it answers with ntpq's
associations as "rows") and so never found a server.
2026-11-11 15:20:29 +01:00
Christian Manivong b50732c63c feat(ssh): say where the SSH algorithm lists are set
driver / lint (pull_request) Successful in 15s
driver / standard (pull_request) Successful in 18s
driver / types (pull_request) Successful in 1m13s
driver / test (3.13) (pull_request) Successful in 1m13s
driver / test (3.12) (pull_request) Successful in 1m15s
driver / lowest (pull_request) Successful in 40s
driver / build (pull_request) Successful in 37s
driver / security (pull_request) Successful in 1m6s
CI / driver (pull_request) Successful in 2m37s
driver / publish (pull_request) Skipped
OPNsense writes sshd_config from config.xml (<system><ssh>), set on the
legacy Administration page; no API controller reads or writes that
section (26.1, 26.7, master). The driver cannot harden the SSH server,
so it declares, as data, the page and the field of each algorithm list:
SSH_SETTINGS_MENU and SSH_SETTINGS_FIELDS, keyed like device-types'
SSH_ALGORITHM_CLASSES. An application tells a person what to set there.

The CARP master check moves to virtual_ips, next to the CARP reader, so
opnsense.py stays under its line ratchet.

Refs NetOrk/netork#1769
2026-11-07 16:39:12 +01:00
Christian Manivong c4ff19ac20 feat(scheduled-jobs): the cron jobs of config.xml over the REST API (#43)
driver / standard (pull_request) Successful in 25s
driver / lint (pull_request) Successful in 26s
driver / types (pull_request) Failing after 35s
driver / test (3.13) (pull_request) Failing after 18s
driver / test (3.12) (pull_request) Failing after 21s
driver / lowest (pull_request) Failing after 17s
driver / build (pull_request) Successful in 30s
driver / security (pull_request) Failing after 40s
CI / driver (pull_request) Failing after 1m25s
driver / publish (pull_request) Skipped
OPNsense keeps its cron jobs in its configuration and runs each as a
configd action, so the POSIX implementation does not fit: the driver has
no shell, and a job's command is not a shell line. The cron API reads and
writes them instead.

get_scheduled_jobs reads every job in one request (/api/cron/settings/get,
which carries the selected action rather than its display text). A job
another component registered is platform, one written by
put_scheduled_job carries the napalm-scheduled-job marker in its
description and is application, the rest admin. Every job can be enabled
and disabled; only an application's job is edited or deleted.
put_scheduled_job takes a configd action as command and refuses one
configd does not offer; get_scheduled_job_commands lists them. Running a
job now is refused: the API cannot run a configd action. Every write
reads the job again and refuses one whose digest changed.

Parameters are written one per line and only without blanks, quotes or
backslashes: OPNsense up to 26.1 splits them as a shell would, 26.7 takes
a line per argument, and only such arguments mean the same to both.

Requires napalm-device-types 4.26.0.

Closes #43
2026-11-03 05:44:29 +01:00
Christian Manivong 72b4806c65 chore: Python 3.12 is the floor (NAPALM/driver-standard#6)
driver / standard (pull_request) Successful in 14s
driver / lint (pull_request) Successful in 14s
driver / types (pull_request) Successful in 52s
driver / lowest (pull_request) Failing after 38s
driver / test (3.12) (pull_request) Successful in 1m18s
driver / test (3.13) (pull_request) Successful in 1m13s
driver / build (pull_request) Successful in 29s
driver / security (pull_request) Successful in 1m4s
CI / driver (pull_request) Failing after 2m49s
driver / publish (pull_request) Skipped
netOrk runs every driver in its own containers on Python 3.12, so 3.10
and 3.11 are no longer supported or tested. requires-python, mypy and
ruff move to 3.12, along with what ruff's pyupgrade rules then rewrite.
2026-10-27 10:18:58 +01:00
Christian Manivong 6e0e206e06 chore: adopt the NAPALM driver standard
driver / standard (pull_request) Successful in 15s
driver / lint (pull_request) Successful in 12s
driver / types (pull_request) Successful in 1m38s
driver / test (3.10) (pull_request) Successful in 1m1s
driver / test (3.11) (pull_request) Successful in 52s
driver / test (3.12) (pull_request) Successful in 49s
driver / test (3.13) (pull_request) Successful in 45s
driver / lowest (pull_request) Successful in 29s
driver / security (pull_request) Successful in 47s
driver / build (pull_request) Successful in 22s
CI / driver (pull_request) Successful in 7m11s
driver / publish (pull_request) Skipped
- CI moves to .gitea/workflows/ci.yml and calls the shared driver workflow
  of NAPALM/driver-standard; .github/ is gone.
- setuptools>=77 with the SPDX license string, no License classifier,
  Python 3.10-3.13 classifiers, napalm-device-types>=4.21.1, a dev extra
  with floors, links to git.netork.io, LICENSE file, py.typed.
- CHANGELOG.md in Keep-a-Changelog form; README names Python 3.10, the
  Gitea clone URL and the new CI.
- mypy strict for the package; exempt for now (52 errors): ha, listening,
  opnsense, ospf_mixin, ping_mixin, port_forwards, resolver_mixin.
  Recorded in the quality baseline with the long and complex functions,
  the 3100-line opnsense.py and one unannotated public method.
- Coverage floor 79 % (measured 79.20 %).
2026-10-23 08:35:15 +02:00
Christian Manivong 571d9abb5f style: ruff format and the lint rules of the driver standard
Mechanical: ruff format, and ruff check --fix with the rule set every NAPALM
driver uses (E, F, I, UP, B at line length 100, Python 3.10). What ruff
cannot fix comes in the next commit.
2026-10-23 08:32:41 +02:00
Christian Manivong e3ed0f1bdd fix: name and version from system_information; a CARP address on its MASTER's interface
CI / test (3.10) (push) Successful in 30s
CI / test (3.11) (push) Successful in 29s
CI / test (3.12) (push) Successful in 30s
CI / test (3.10) (pull_request) Successful in 30s
CI / test (3.11) (pull_request) Successful in 28s
CI / test (3.12) (pull_request) Successful in 29s
get_facts read hostname and version from /api/core/system/status, which on
OPNsense 26.7 carries neither: every firewall read as nameless and "unknown",
and netOrk kept naming one by its address. /api/diagnostics/system/system_information
has both (name is the FQDN, versions[0] "OPNsense 26.7.7-amd64"). The status
answer still counts where the new one gives nothing.

get_interfaces_ip listed each interface's primary address only, so a CARP address
never showed as the holder's. It now lists an address bound to a VHID on the node
that is MASTER for it, and leaves it off a BACKUP -- as keepalived puts the address
on the master alone. netOrk's VIP badge sits on that interface (netOrk #1121).
2026-10-16 15:37:43 +02:00
Christian Manivong 00172f7a69 feat: report the firewall's CARP addresses as virtual IPs
CI / test (3.10) (push) Successful in 30s
CI / test (3.11) (push) Successful in 29s
CI / test (3.12) (push) Successful in 30s
CI / test (3.10) (pull_request) Successful in 29s
CI / test (3.11) (pull_request) Successful in 28s
CI / test (3.12) (pull_request) Successful in 30s
get_virtual_ips() reads every address bound to a CARP VHID -- the CARP
address and IP aliases on its VHID alike -- from
/api/interfaces/overview/export, with this node's CARP state, so the reading
says who holds each address. interfaces/vip_settings would list the same
addresses with each VHID's CARP password; it is not read.

vip_manager and keepalived are Linux mechanisms and are None here. Needs
napalm-device-types 4.10.0 (carp_virtual_ip).

netOrk #1121.
2026-10-16 11:07:08 +02:00
Christian Manivong 2fb1226734 feat: report what listens, and which OPNsense service it is
CI / test (3.10) (push) Successful in 40s
CI / test (3.11) (push) Successful in 30s
CI / test (3.12) (push) Successful in 31s
CI / test (3.10) (pull_request) Successful in 29s
CI / test (3.11) (pull_request) Successful in 28s
CI / test (3.12) (pull_request) Successful in 30s
get_listening_sockets reads /api/diagnostics/interface/get_socket_statistics:
netstat's sockets with sockstat's user, command and PID, collected by
configd as root, so no SSH is needed. A listening socket is one without a
peer (*:*); "*" is the any-address of the socket's family.

Which service: the socket's unit is a name of the firewall's own service
list, so netOrk can match it to the service. The command (cut to ten
characters by FreeBSD) matches a service name, the start of one, or one
of the daemons whose service is called otherwise (sshd is openssh, the
FRR daemons are frr, kea-ctrl-agent is kea-dhcp). lighttpd is the web UI,
or the captive portal when it runs as www. WireGuard's sockets belong to
the kernel and are told by the listen ports of /api/wireguard/service/show.

The fixture is a real OPNsense 26.7.5 firewall's answer, cut down, with
documentation addresses. For netOrk#673.
2026-10-07 07:20:27 +02:00
Christian Manivong f934cc0cfa feat(ping): implement ping and a batched ping_sweep over the diagnostics API
CI / test (3.10) (push) Failing after 19s
CI / test (3.11) (push) Failing after 8s
CI / test (3.12) (push) Failing after 11s
CI / test (3.9) (push) Failing after 8s
OPNsense has no synchronous ping endpoint. /api/diagnostics/ping is a job API —
create, start, read statistics, stop, remove — so a single ping costs five
requests and roughly a second of waiting, which makes the generic per-host
sweep from napalm-device-types unusable for a whole subnet.

The override exploits what the job API does offer instead: jobs are
independent and run on the firewall in parallel, and search_jobs reports all of
them in one response. A batch (32 by default) is created and started, waited
for once, harvested with a single request and then cleaned up, so waiting time
per batch is constant rather than linear in hosts. PING_SWEEP_MAX_TARGETS
bounds the sweep as a whole — this runs on production firewalls.

Two details the API forces: results are polled, because search_jobs signals the
running ping with SIGINFO and then parses whatever it has written so far, so
the first read of a healthy host can still show zero probes; and the model's
root node is read from /get rather than hardcoded, so a rename in a future
OPNsense release cannot silently break job creation.

Tested against mocked API responses only — no live device is reachable at the
moment, so the endpoint shapes come from the OPNsense sources (PingController,
scripts/interfaces/ping.py).
2026-08-13 16:51:05 +07:00
Christian Manivong ae27cd5469 initial commit 2026-05-29 09:22:10 +02:00