Server data
Extensions to NETCONF let a client ask the server about itself: its sessions and counters, its YANG modules and their source, its datastores. Each is a command when the server has the capability for it, and help lists the rest under "not available". These commands only read, so they work on read-only targets too.
Monitoring
With RFC 6022 monitoring, the server reports its open sessions, the schemas it offers, its datastores and who has them locked, and its counters. nnc prints each as JSON:
nnc lab> monitoring-sessions
nnc lab> monitoring-statistics
nnc lab> monitoring-datastores
nnc lab> monitoring-schemasnnc lab> monitoring-sessions
[
{
"sessionId": 3,
"transport": "ietf-netconf-monitoring:netconf-ssh",
"username": "netconf",
"sourceHost": "192.0.2.50",
"loginTime": "2026-11-14T10:00:32Z",
"inRpcs": 41,
"inBadRpcs": 0,
"outRpcErrors": 2,
"outNotifications": 0
}
]monitoring-datastores shows a lock's holder and when it was taken, which is how to find the session to kill-session:
nnc lab> lock running
nnc lab> monitoring-datastores
nnc lab> unlock runningFetching a module
get-schema fetches a module's source from the server, as YANG or YIN:
nnc lab> get-schema ietf-system
nnc lab> get-schema ietf-system --format yinget-schema ietf-system > ietf-system.yang saves it to a file, to read or to compile your own tools against.
The YANG library
A server with a YANG library (RFC 8525, or RFC 7895's /modules-state) lists the modules each datastore implements, with their features and deviations. modules shows it as a table, and yang-library in full:
nnc lab> modules
nnc lab> yang-librarynnc reads the YANG library once per session. Its content-id, which the server also gives in its hello, changes when the modules do; yang-library --refresh reads it again.
NMDA datastores
A server with the Network Management Datastore Architecture (RFC 8342) has the intended and operational datastores as well as running and candidate. datastores lists them, and get-data reads any of them (RFC 8526):
nnc lab> datastores
nnc lab> get-data intended --filter subtree {"ietf-system:system": {}}
nnc lab> get-data operational --filter xpath /ietf-interfaces:interfacesoperational holds the configuration in use and the state data together.
| Option | |
|---|---|
--config-filter / --no-config-filter | configuration nodes only, or state nodes only |
--max-depth <n> | levels below each selected node |
--with-origin | where each configuration value came from (operational only) |
--origin-filter <origin>, --negated-origin-filter <origin> | keep, or drop, nodes with that origin |
--with-defaults <mode> | as for get-config |
edit-data changes an NMDA datastore, and always rolls back if any part fails:
nnc lab> edit-data running {"ietf-system:system": {"location": "rack-1"}}
nnc lab> get-data running --filter subtree {"ietf-system:system": {}}Origins
--with-origin shows where each value in the operational datastore came from: intended (your configuration), learned (a protocol, such as LLDP), system, default, dynamic or unknown. nnc shows it as one line per leaf, including the origins leaves inherit from their parents:
nnc lab> get-data operational --filter xpath /ietf-system:system --with-origin/ietf-system:system/contact tutorial@example.net intended
/ietf-system:system/hostname lab1 intended
2 leaves: 2 intended--json gives [{ "path", "value", "origin" }], and --xml the reply with its or:origin attributes.
Comparing on the server
A server with RFC 9144 compares two NMDA datastores itself. compare shows the differences as a short diff:
nnc router1> compare running operational
- delete /ietf-interfaces:interfaces/interface[name=eth1]
~ replace /ietf-interfaces:interfaces/interface[name=eth0]/mtu 1500 → 9000
+ create /ietf-interfaces:interfaces/interface[name=eth2]
3 differences (running → operational)--all compares every node, not only those both datastores can hold, --report-origin adds the origins from operational, and --filter compares part of them. On any server, nnc's own diff compares running, candidate and startup.
Partial locks
RFC 5717 partial locks lock part of the running configuration, chosen by XPath, instead of all of it:
nnc router1> partial-lock /ietf-interfaces:interfaces/interface[name='eth0']
nnc router1> partial-unlock 1A part already locked by another session is refused, naming the session. Few servers implement it; the lab device doesn't.
Next
- Device profiles: what nnc knows about a kind of device
- Notifications: subscriptions, and the streams they read