Commit Graph
9 Commits
Author SHA1 Message Date
christianmanivong ae09e44345 Merge pull request 'fix: raise when an update reader cannot read, and report host status' (#5) from feat/host-status into main
CI / test (3.10) (push) Successful in 25s
CI / test (3.11) (push) Successful in 23s
CI / test (3.12) (push) Successful in 27s
2026-10-08 10:15:13 +00:00
Christian Manivong 25308bf747 fix: raise when an update reader cannot read, and report host status
CI / test (3.10) (push) Successful in 26s
CI / test (3.11) (push) Successful in 25s
CI / test (3.12) (push) Successful in 26s
CI / test (3.10) (pull_request) Successful in 25s
CI / test (3.11) (pull_request) Successful in 24s
CI / test (3.12) (pull_request) Successful in 26s
netOrk reads an empty update list as "no updates" and closes every patch
clock on the host. A refused sudo, a failing pkg, syspatch or freebsd-update,
or a pkg database pkg cannot read used to come back as that empty list. Each
of these now raises. Only pkg's "is not installed" still means no packages.

BsdDriver takes on napalm-device-types' HostStatusMixin. On FreeBSD it
reports a host whose installed kernel differs from the running one as
needing a reboot (device-types 4.1). OpenBSD stays unknown.

Closes #4
2026-10-08 12:08:44 +02:00
christianmanivong 1172b4d9d5 Merge pull request 'feat: set up SNMP with net-snmp' (#3) from feat/snmp into main
CI / test (3.10) (push) Successful in 27s
CI / test (3.11) (push) Successful in 26s
CI / test (3.12) (push) Successful in 28s
2026-10-08 08:08:46 +00:00
christianmanivong aef58ca161 feat: set up SNMP with net-snmp
CI / test (3.10) (pull_request) Successful in 30s
CI / test (3.11) (pull_request) Successful in 32s
CI / test (3.12) (pull_request) Successful in 36s
CI / test (3.10) (push) Successful in 45s
CI / test (3.11) (push) Successful in 30s
CI / test (3.12) (push) Successful in 29s
run_device_action("fix_snmp") and get_snmp_config for both drivers, as
decided in NetOrk/netork#800: net-snmp from packages rather than the base
daemons, because it answers UCD-SNMP-MIB, which netOrk's health metrics
read.

- fix_snmp installs net-snmp, writes the configuration netOrk uses on
  Linux (v2c, community "public", agentAddress udp:161) through stdin as
  root, enables and restarts the agent (FreeBSD `service snmpd`, OpenBSD
  `rcctl ... netsnmpd`, as `snmpd` is OpenBSD's own daemon) and asks it
  for sysDescr. It stops at the first step that fails and says why.
- get_snmp_config reports a running net-snmp's community and port.
- FreeBSD's install_package bootstraps pkg on a classic system that never
  had it instead of waiting for an answer that never comes.

Checked live: FreeBSD from a bare system, OpenBSD again over an existing
setup. Memory, swap and load answer on both; FreeBSD's ssCpuIdle about a
minute after start; OpenBSD's CPU figures from net-snmp are wrong (#800).
2026-10-08 10:05:09 +02:00
christianmanivong 7bf028ef4d Merge pull request 'feat: packages, services and updates' (#2) from feat/packages-services-updates into main
CI / test (3.10) (push) Successful in 28s
CI / test (3.11) (push) Successful in 26s
CI / test (3.12) (push) Successful in 28s
2026-10-08 08:02:25 +00:00
christianmanivong e8468a292a feat: packages, services and updates
CI / test (3.10) (push) Successful in 26s
CI / test (3.11) (push) Successful in 24s
CI / test (3.12) (push) Successful in 26s
CI / test (3.10) (pull_request) Successful in 25s
CI / test (3.11) (pull_request) Successful in 24s
CI / test (3.12) (pull_request) Successful in 25s
The write side NetOrk/netork#799 needs, on both drivers.

- Packages: get_packages (FreeBSD `pkg query` with the repository as
  source, nothing on a classic system without pkg; OpenBSD `pkg_info`),
  install_package / uninstall_package as root (`pkg install`/`pkg
  delete`, `pkg_add -I`/`pkg_delete`), names checked before anything is
  sent, a failure raised with what the tool printed.
- Updates: get_available_updates. FreeBSD `pkg upgrade -n`, with
  `security` from VuXML (`pkg audit`: True for a listed package, False for
  any other, None when the audit could not run); OpenBSD `pkg_add -u -n
  -v` (no security verdict). Base-system patches are one `base-system`
  entry, as decided in #799: OpenBSD's `syspatch -c`, and on a classic
  FreeBSD what `freebsd-update` has fetched; on pkgbase the base system is
  packages from FreeBSD-base.
- Services: get_services lists the enabled ones (FreeBSD `service ...
  status` as root, as root-only pidfiles hide daemons otherwise, without
  root when sudo refuses; OpenBSD `rcctl check`), manage_service runs
  start/stop/restart/enable/disable as root and returns success and
  output.

Fixtures recorded on the FreeBSD 15.1 and OpenBSD 7.9 VMs with genuinely
outdated packages (FreeBSD pointed at the latest branch, an OpenBSD
package taken back to its release build). Checked live on both, including
a service restart and installing and removing a package.
2026-10-08 09:59:04 +02:00
christianmanivong 045f809602 Merge pull request 'feat: report listening sockets' (#1) from feat/listening-sockets into main
CI / test (3.10) (push) Successful in 28s
CI / test (3.11) (push) Successful in 28s
CI / test (3.12) (push) Successful in 31s
2026-10-08 07:30:25 +00:00
christianmanivong 5f53de0c25 feat: report listening sockets
CI / test (3.10) (push) Successful in 27s
CI / test (3.11) (push) Successful in 27s
CI / test (3.12) (push) Successful in 37s
CI / test (3.10) (pull_request) Successful in 42s
CI / test (3.11) (pull_request) Successful in 42s
CI / test (3.12) (pull_request) Successful in 38s
get_listening_sockets() in the shape of napalm-device-types'
ListeningSocketsMixin, whose ss/cgroup reading is Linux's, and with its
rule: read as root first, and without root when that brings nothing back,
which the reading reports as `attributed: False`.

- FreeBSD: `sockstat -46lq -P tcp,udp`, which sees every socket either way.
- OpenBSD: `fstat -n` as root (the sockets nothing is connected to; port 0
  is unbound), `netstat -an` without root, as fstat shows a user only
  their own processes; those sockets come without a process.
- Addresses as ss prints them: `*` becomes 0.0.0.0 or :: by family, IPv6
  without brackets, a zone (`fe80::1%lo0`) becomes the interface. One entry
  per socket, the first process holding it.

No get_kernel_facts on purpose: KernelFactsMixin reports a Linux kernel's
modules and CONFIG_ options, which a BSD kernel does not have.

Fixtures recorded on the FreeBSD 15.1 and OpenBSD 7.9 VMs with test
listeners on 127.0.0.1 and ::1 (NetOrk/netork#798).
2026-10-08 09:26:21 +02:00
christianmanivong 14afd33798 feat: FreeBSD and OpenBSD drivers, read side
CI / test (3.10) (push) Successful in 42s
CI / test (3.11) (push) Successful in 37s
CI / test (3.12) (push) Successful in 29s
FreeBSDDriver (`freebsd`) and OpenBSDDriver (`openbsd`) on a shared
BsdDriver, for the BSD VMs netOrk provisions and BSD hosts added by hand
(NetOrk/netork#798).

- SSH over exec channels (napalm-device-types' run_on_transport): no PTY,
  stdout and stderr apart, real exit codes; run_command()/open_stream()
  are public. Root through `sudo -S` with the password on stdin, or
  `sudo -n` without one.
- Facts: hostname, release and running kernel (freebsd-version / uname),
  hardware from SMBIOS (kenv) or hw.* sysctls, uptime from kern.boottime.
- Interfaces and addresses (ifconfig -a: ether/lladdr, hex netmasks,
  scoped IPv6), routes (netstat -rn, OpenBSD's abbreviated networks
  written out, host/broadcast/loopback entries left out), ARP (FreeBSD's
  sentences, OpenBSD's table), users and groups, processes (lstart read as
  one five-word field), cron jobs (system crontab and the login user's).
- Parsers are pure functions, tested on output recorded from FreeBSD 15.1
  (hand-installed and cloud image) and OpenBSD 7.9 (NetOrk/netork#793),
  with internal addresses replaced by documentation ranges.
2026-10-08 09:12:53 +02:00