The project provides a command-line facilities for remotely
interacting with the administrative and diagnostic services of the onos-config
server.
The commands are available at run-time using the consolidated onos
client hosted in
the onos-cli
repository, but their implementation is hosted and built here.
The documentation about building and deploying the consolidate onos
client or its Docker container
is available in the onos-cli
GitHub repository.
> onos config --help
ONOS configuration subsystem commands
Usage:
onos config [command]
Available Commands:
add Add a config resource
compact-changes Takes a snapshot of network and device changes
config Manage the CLI configuration
get Get config resources
load Load configuration from a file
rollback Rolls-back a network change
snapshot Commands for managing snapshots
watch Watch for updates to a config resource type
Flags:
-h, --help help for config
--no-tls if present, do not use TLS
--service-address string the onos-config service address (default "onos-config:5150")
--tls-cert-path string the path to the TLS certificate
--tls-key-path string the path to the TLS key
Use "onos config [command] --help" for more information about a command.
Since the onos
command is a client, it requires the address of the server as well
as the paths to the key and the certificate to establish secure connection to the
server.
These options are global to all commands and can be persisted to avoid having to specify them for each command. For example, you can set the default server address as follows:
> onos config config set address onos-config-server:5150
Subsequent usages of the onos
command can then abstain from using the --address
option to indicate the server address, resulting in easier usage.
To rollback a network use the rollback admin tool. This will rollback the last network
change unless a specific change is given with the changename
parameter
> onos config rollback Change-VgUAZI928B644v/2XQ0n24x0SjA=
A model plugin is a shared object library that represents the YANG models of a particular Device Type and Version. The plugin allows user to create and load their own device models in to onos-config that can be used for validating that configuration changes observe the structure of the YANG models in use on the device. This improves usability by pushing information about the devices' model back up to the onos-config gNMI northbound interface.
Model plugins can be loaded at the startup of onos-config by (repeated) --modelPlugin
options, or they can be loaded at run time. To see the list of currently loaded
plugins use the command:
> onos config get plugins
There are a number of commands that provide internal view into the state the onos-config store. These tools use a special-purpose gRPC interfaces to obtain the internal meta-data from the running onos-config process. Please note that these tools are intended purely for diagnostics and should not be relied upon for programmatic purposes as they are not subject to any backward compatibility guarantees.
For example, to list and watch all changes stored internally run:
> onos config watch network-changes
...
or to watch device-changes
> onos config watch device-changes
...
Configuration data can be loaded in to onos-config through the cli with
onos config load yaml <filename(s)>
The Yaml file must be in the form below. Several updates, replace or delete entries can be made.
This effectively is the same as a gNMI SetRequest, but with the input in YAML format instead of PROTO.
Only the SeqRequest functionality is possible with this command. To do a gNMI GetRequest use the gnmi_cli tool [GetRequest](./gnmi.md#Northbound gNMI Get Request)
A separate NetworkChange will be created for each file given.
This allows the set of updates to be broken up in to smaller groups.
setrequest:
prefix:
elem:
- name: e2node
target: ""
delete: []
replace: []
update:
- path:
elem:
- name: intervals
- name: RadioMeasReportPerUe
target: 315010-0001420
val:
uintvalue:
uintval: 20
duplicates: 0
extension:
- id: 101
value: 1.0.0
- id: 102
value: E2Node