You can build the locally by executing
go build ./...
You can build the docker container by running:
docker build .
NSM_NAME
- Name of EndpointNSM_LABELS
- Labels related to this forwarder-vpp instanceNSM_NSNAME
- Name of Network Service to Register with RegistryNSM_CONNECT_TO
- url to connect toNSM_LISTEN_ON
- url to listen onNSM_MAX_TOKEN_LIFETIME
- maximum lifetime of tokensNSM_REGISTRY_CLIENT_POLICIES
- paths to files and directories that contain registry client policiesNSM_LOG_LEVEL
- Log levelNSM_DIAL_TIMEOUT
- Timeout for the dial the next endpointNSM_OPEN_TELEMETRY_ENDPOINT
- OpenTelemetry Collector EndpointNSM_METRICS_EXPORT_INTERVAL
- interval between mertics exportsNSM_PPROF_ENABLED
- is pprof enabled (default: "false")NSM_PPROF_LISTEN_ON
- pprof URL to ListenAndServe (default: "localhost:6060")NSM_PROMETHEUS_LISTEN_ON
- Prometheus URL to ListenAndServe (default: ":8081")NSM_PROMETHEUS_SERVER_HEADER_TIMEOUT
- sets the header timeout for the Prometheus metrics server (default: "5s")NSM_TUNNEL_IP
- IP to use for tunnelsNSM_VXLAN_PORT
- VXLAN port to useNSM_VPP_API_SOCKET
- filename of socket to connect to existing VPP instance.NSM_VPP_INIT
- type of VPP initialization. Must be AF_XDP, AF_PACKET or NONENSM_VPP_INIT_PARAMS
- Configuration file path containing VPP API parameters for initializationNSM_VPP_MIN_OPERATION_TIMEOUT
- minimum timeout for every vpp operationNSM_RESOURCE_POLL_TIMEOUT
- device plugin polling timeoutNSM_DEVICE_PLUGIN_PATH
- path to the device plugin directoryNSM_POD_RESOURCES_PATH
- path to the pod resources directoryNSM_DEVICE_SELECTOR_FILE
- config file for device name to label matchingNSM_SRIOV_CONFIG_FILE
- PCI resources config pathNSM_PCI_DEVICES_PATH
- path to the PCI devices directoryNSM_PCI_DRIVERS_PATH
- path to the PCI drivers directoryNSM_CGROUP_PATH
- path to the host cgroup directoryNSM_VFIO_PATH
- path to the host VFIO directoryNSM_MECHANISM_PRIORITY
- sets priorities for mechanisms
Testing is run via a Docker container. To run testing run:
docker run --privileged --rm $(docker build -q --target test .)
If you wish to debug the test code itself, that can be acheived by running:
docker run --privileged --rm -p 40000:40000 $(docker build -q --target debug .)
This will result in the tests running under dlv. Connecting your debugger to localhost:40000 will allow you to debug.
-p 40000:40000
forwards port 40000 in the container to localhost:40000 where you can attach with your debugger.
--target debug
Runs the debug target, which is just like the test target, but starts tests with dlv listening on port 40000 inside the container.
When you run 'cmd' you will see an early line of output that tells you:
Setting env variable DLV_LISTEN_FORWARDER to a valid dlv '--listen' value will cause the dlv debugger to execute this binary and listen as directed.
If you follow those instructions when running the Docker container:
docker run --privileged -e DLV_LISTEN_FORWARDER=:50000 -p 50000:50000 --rm $(docker build -q --target test .)
-e DLV_LISTEN_FORWARDER=:50000
tells docker to set the environment variable DLV_LISTEN_FORWARDER to :50000 telling
dlv to listen on port 50000.
-p 50000:50000
tells docker to forward port 50000 in the container to port 50000 in the host. From there, you can
just connect dlv using your favorite IDE and debug cmd.
docker run --privileged -e DLV_LISTEN_FORWARDER=:50000 -p 40000:40000 -p 50000:50000 --rm $(docker build -q --target debug .)
Please note, the tests start the cmd, so until you connect to port 40000 with your debugger and walk the tests through to the point of running cmd, you will not be able to attach a debugger on port 50000 to the cmd.
Because cmd-forwarder-vpp is only anticipated to run in Linux, you will need to run golangci-lint run with:
GOOS=linux golangci-lint run