Running mock switches¶

ngsw serve runs the mocks as long-lived daemons on real sockets, so any tool — in any language — can talk to one when hardware is unavailable.

ngsw serve --model gsm7228ps                     # one model
ngsw serve --model gsm7228ps --model gs305ep     # several
ngsw serve --all                                 # every registered model

Each switch that comes up prints its model, bind address, the port it actually bound for each protocol, and the credentials it accepts. The command then blocks until interrupted, stopping every switch cleanly.

[m4300-24x] host=127.0.0.1
    SNMP udp/46382
    HTTP tcp/41507
    community='public' http_password='password'
[gs110emx] host=127.0.0.1
    NSDP udp/54443
    HTTP tcp/40875
    community='public' http_password='password'
[m7300] host=127.0.0.1
    SNMP udp/55530
    community='public' http_password='password'
serving 10 mock switch(es); press Ctrl-C to stop

The model’s registry entry decides which faces bind, which is why the GS110EMX above offers NSDP and HTTP while the M7300 offers only SNMP — the mocks have exactly the backends the real switches have.

Options¶

Option	Meaning
`--model KEY`	Model to serve; repeatable. `ngsw models` lists the keys.
`--all`	Serve every registered model.
`--host IP`	Bind address, default `127.0.0.1`. Use `0.0.0.0` to expose the mock to other hosts.
`--community STR`	SNMP community the mock accepts, default `public`.
`--http-password STR`	Web-UI password the mock accepts, default `password`.
`--port N` / `--http-port N`	Pin the UDP (SNMP or NSDP) and TCP (HTTP) ports instead of using ephemeral ones.

Note

A pinned port is a single listener, so --port/--http-port are only accepted when serving exactly one model. When serving several, leave them unset and read the printed ports.

One bad model never takes the fleet down: a switch that cannot bind is reported and skipped, and the rest keep serving.

Pointing tools at it¶

SNMP¶

Any SNMP client works, because the face is a real agent on a real UDP socket:

ngsw serve --model gsm7228ps --port 16161 &

snmpget -v2c -c public 127.0.0.1:16161 1.3.6.1.2.1.1.1.0
snmpwalk -v2c -c public 127.0.0.1:16161 1.3.6.1.2.1.1
snmpbulkwalk -v2c -c public 127.0.0.1:16161 1.3.6.1.2.1.2.2.1.8

iso.3.6.1.2.1.1.1.0 = STRING: "S3300-52X-PoE+ ProSAFE 48-Port Gigabit
Stackable Smart Switch with PoE+ and 4 10G uplinks"
iso.3.6.1.2.1.1.2.0 = OID: iso.3.6.1.4.1.4526.100.10.19

That sysDescr is the real switch’s, from a capture — which is what makes the mock useful for testing model detection, not just data parsing.

SNMP writes work too, and so do the refusals:

snmpset -v2c -c public 127.0.0.1:16161 1.3.6.1.2.1.2.2.1.7.1 i 2

NSDP¶

The NSDP face answers on the printed UDP port. Because NSDP is normally broadcast to port 63322, a client usually needs to be told where to send; this library’s own transport takes the port directly, and third-party tools have been driven against it over a veth pair in a network namespace.

HTTP¶

The HTTP face serves the model’s real page set, so curl reaches it directly:

ngsw serve --model gs305ep --http-port 18080 &
curl -s http://127.0.0.1:18080/dashboard.cgi

<html><body><input type="hidden" name="hash" value="virtualhash"><table>
<tr class="portID"><td><input type="checkbox"></td><td>1</td><td>Up 1000M</td>
<td>Enabled</td><td>Port 1</td></tr>...

The login endpoint is served too, and the mock validates the posted password. Login paths and fields differ per model — the GS305EP posts to /login.cgi while the S3300 uses /base/cheetah_login.html — so a tool that hard-codes one model’s flow will not work against another, exactly as with real switches.

Warning

Known deviation. The HTTP face validates the login POST, and reproduces the M4300’s Referer-header CSRF check (answering 403 without it), but it does not require a prior session for ordinary page GETs. Real firmware does. If what you are testing is your tool’s session handling — login, expiry, re-login — the mock will be more permissive than the device. Everything else about the pages is faithful.

The FASTPATH CLI¶

The CLI face is deliberately in-process: it implements the CliSession protocol rather than binding an SSH or telnet listener, so there is no key exchange or terminal emulation to fight. Reach it through cli_session, not over a socket; ngsw serve therefore does not expose it. See Testing against the fake.

This library against a mock¶

ngsw itself is a perfectly good external tool:

ngsw serve --model gsm7228ps --port 16161 &
ngsw --host '127.0.0.1:16161' --model gsm7228ps --community public ports

Running one in CI¶

Start it in the background, wait for the port, run your tests, stop it:

# .github/workflows/test.yml
- name: Start a mock Netgear switch
  run: |
    pip install python-netgear-switch-library
    ngsw serve --model gsm7228ps --port 16161 --http-port 18080 &
    for _ in $(seq 1 50); do
      nc -z -u 127.0.0.1 16161 && break
      sleep 0.2
    done

- name: Test against it
  run: pytest
  env:
    SWITCH_HOST: 127.0.0.1:16161

For pytest specifically, an in-process fixture is simpler and faster than a subprocess — see Testing against the fake.