Skip to content

Latest commit

 

History

History
75 lines (56 loc) · 3.75 KB

setup.md

File metadata and controls

75 lines (56 loc) · 3.75 KB
Raw

Development Setup

This section describes necessary steps to setup a development environment for testing and contributing to this project.

Requirements

  • go >= 1.20
  • git, make and kubectl
  • Access to a Kubernetes cluster (Minikube, kind or a real cluster). We also provide detailed guidance on setting up a kind cluster for local development.
  • metalbond We need to setup a metalbond server that is responsible of managing overlay routes in a network. Metalnet establishes a connection with it during the initialization.
  • dp-service We need to start a dp-service process that reacts to the gRPC calls generated from metalbond.

Metalbond and dp-service

Assuming the executable binaries of metalbond and dp-service has been successfully compiled by following their corresponding instructions, we first need to start these two components. For more detailed information on these two components, please refer to their documentation.

To start metalbond server, we use the following command. Switch to metalbond's repo and run

./metalbond server --listen [::]:4711 --http [::1]:4712 --keepalive 3

Additionally, the default router address can be announced by running the following command. In this case, no extra parameter needs to be provided when starting the controller manager.

 ./metalbond client      --server "[::]:4711"      --keepalive 3      --subscribe 100      --announce 100#0.0.0.0/0#[default-router-address, e.g., abcd:efgh:1234::5]

To start dp-service, we use the following command. Switch to dp-service's repo and run

sudo ./build/src/dpservice-bin -l 0,1  -a [pci-address-of-1st-pf],representor=[vf-in-use, e.g., 0-4]  -a [pci-address-of-2nd-pf] --proc-type=primary --log-level user1:8  -- --pf0=[interface-name-of-1st-pf] --pf1=[interface-name-of-2nd-pf] --vf-pattern=[interface-name-patternof-vfs, e.g., ens1f0npf0vf] --ipv6=[ipv6-addr-of-host] --no-stats --no-offload

Running on the cluster

Before running metalnet operators, it is necessary to ensure that proper access to a kubernetes cluster is obtained. Your controller will automatically use the current context in your kubeconfig file (i.e. whatever cluster kubectl cluster-info shows).

Install Instances of Custom Resources:

Your Kubernetes API server needs to know about the APIs which come with the metalnet project. To install the APIs your cluster, run

make install

Start the Controller Manager

The controller manager can be started via the following command

make run-base

If you want to manually specify the default router address, start the controller manager by running:

go run ./main.go --router-address=[default-router-address]

Apply Sample Manifests

The config/samples folder contains samples for all APIs supported by this project. You can apply any of the samples by running

kubectl apply -f config/samples/SOME_RESOURCE.yaml

Note: It is possible that you need to adapt these manifest samples to your local environment. For example, the nodeName field needs to be changed to the actual worker node's name in your kubernetes cluster.

Rebuilding API Type and Manifests

Everytime a change has been done to any of the types definitions, the corresponding manifests and generated code pieces have to be rebuilt. Make sure your APIs are up-to-date by running make install after your code / manifests have been regenerated.

make generate
make manifests

NOTE: Run make --help for more information on all potential make targets

More information can be found via the Kubebuilder Documentation