gVPC operations

This page covers day-to-day operations on gVPC networks with the network plugin for datumctl. The plugin manages networks, network services, and the network interfaces that Instances hold on your networks. It writes the same resources as the manifests on Configuration, and its describe output explains what’s wrong and who needs to act.

gVPC is in preview, and the networking.datumapis.com/v1alpha API can change.

Install the plugin

To install the plugin from the official Datum catalog, run the following command:

datumctl plugin install network

To check the installation, run the following command:

datumctl network version

datumctl network version works without a login, a project, or a network connection. For more information about plugins, see Using plugins. Without the plugin, you can still list and delete networks with datumctl get networks and datumctl delete network NETWORK_NAME.

Manage networks

The following table describes the network commands:

Command	What it does
`datumctl network create NAME`	Creates an IPv6-only network. Pass `--mtu` to change the MTU from the default of `1440`, and `--wait` to wait until the network is ready.
`datumctl network list`	Lists networks, the locations they’ve reached, and how many Instances and services use them. Pass `-o wide` for each network’s IPv6 prefix.
`datumctl network describe NAME`	Shows a network’s address families, MTU, IPv6 prefix, each location with its address range and Instance counts, the workloads and network services on it, and a diagnosis.
`datumctl network update NAME --mtu MTU`	Changes the MTU, which is the only setting that you can change after you create a network. The change applies to every Instance on the network.
`datumctl network delete NAME`	Deletes a network.

A new network isn’t in any location. It reaches a location when a workload in that location attaches to it, so describe shows no locations until you deploy a workload:

datumctl network create backend
datumctl compute deploy api --image=ghcr.io/example/api:1.4.2 --location=us-east-1 --network=backend
datumctl network describe backend

A workload’s network can’t change after you create the workload. To move a workload to another network, deploy a new workload on that network and then delete the old one. For more information, see Compute operations.

Delete a network

Before you delete a network, delete the workloads that are attached to it. To delete a network, run the following command:

datumctl network delete NETWORK_NAME

Replace NETWORK_NAME with the name of the network. Before it asks you to confirm, the command lists the workloads that lose their network, the network services that stop serving, and the kept interfaces whose addresses are released. If Instances still hold addresses on the network, the network stays in deletion and reports RangeOccupied until they’re gone.

Manage network services

The datumctl network service commands create, describe, change, and delete network services. For more information, see Network services.

datumctl network service create api --workload=api --port=http:8080
datumctl network service list
datumctl network service describe api

Manage network interfaces

Each Instance holds a network interface on its workload’s network. To list the interfaces in your project, run the following command:

datumctl network interface list

To narrow the list, pass --network, --workload, or --kept. A workload with reclaimPolicy: Retain keeps an Instance’s addresses after the Instance is removed. The kept interface holds the addresses until an Instance with the same name returns. For more information, see Keep addresses after scale-down. To release a kept interface and give up its addresses permanently, run the following command:

datumctl network interface release INTERFACE_NAME

Replace INTERFACE_NAME with the name of the kept interface. You can’t release an interface that a running Instance holds. After you release the addresses, update anything that refers to them, such as DNS records or firewall rules.

Find out what’s wrong

datumctl network describe and datumctl network service describe end with a diagnosis that names the cause, the location or workload that it affects, and who needs to act: you, Datum, or nobody while a change is still in progress. Keep the following points in mind:

A new network with no locations isn’t a fault. The network reaches a location when a workload there attaches to it.
A network service with no ready Instances is a workload problem, not a network problem. datumctl network service describe lists the Instances that aren’t ready and what each one reports.

For specific errors, see Troubleshooting.

Scripts and automation

Confirmation: delete, update, service delete, and interface release ask you to type the resource’s name. Without a terminal, they stop unless you pass --yes.
Previews: Commands that change resources accept --dry-run, which validates the change and discards it.
Output: Commands that read resources accept -o table, -o wide, -o json, -o yaml, or -o name.
Exit codes: Failures print a named exit code, such as exit status 4 # NETWORK_NOT_FOUND. The codes are NETWORK_USAGE (2), NETWORK_FORBIDDEN (3), NETWORK_NOT_FOUND (4), NETWORK_CONFLICT (5), NETWORK_INVALID (6), NETWORK_UNAVAILABLE (8), and NETWORK_ABORTED (9).
Waiting: If create --wait times out, the command exits with an error, but the network exists. Don’t run create again; run describe instead.

What’s next

To create a network with a manifest and attach a workload, see Configuration.
To give a workload a stable name for a load balancer, see Network services.
To fix network errors, see Troubleshooting.

Introduction

Explore

Install the plugin

Manage networks

Delete a network

Manage network services

Manage network interfaces

Find out what’s wrong

Scripts and automation

What’s next

​Install the plugin

​Manage networks

​Delete a network

​Manage network services

​Manage network interfaces

​Find out what’s wrong

​Scripts and automation

​What’s next

Install the plugin

Manage networks

Delete a network

Manage network services

Manage network interfaces

Find out what’s wrong

Scripts and automation

What’s next