Networks

Every instance needs a network interface to reach other instances and the internet. This page explains the network your instances get by default, where to see their addresses, and how to manage networks of your own through the API.

The default network

On a default installation, each tenant's default profile gives every instance an eth0 interface on a shared network that the operator set up. You do not need to create anything: start an instance and it gets an address.

Your operator can instead enable per-tenant networks, where each tenant has its own isolated network space and creates its own networks. Whether this is available depends on the deployment. Ask your operator before you plan around custom networks.

See an instance's network

Open the instance and select the Network tab.

• Interfaces shows the live network state reported by the running instance: each interface's State, Addresses, MAC, MTU, and Received and Sent traffic. It is empty while the instance is stopped.
• NIC devices lists the network devices in force, including ones inherited from profiles, with the network each one is Attached to.

The Overview tab also lists the instance's routable IP addresses in one place.

For scripts, GET /api/v1/compute/instances/{instanceId}/runtime returns the same data. The networks field maps each interface name to its addresses and counters.

Create a network

Networks are managed through the API only. POST /api/v1/compute/networks creates one in your tenant. Only name is required (1 to 63 characters, unique in the tenant); type defaults to bridge.

curl -X POST https://lahijan.example.com/api/v1/compute/networks \  -H "Authorization: Bearer $LAHIJAN_TOKEN" \  -H "X-Tenant-Id: $TENANT_ID" \  -H "Content-Type: application/json" \  -d '{    "name": "backend",    "description": "Private network for the app tier",    "type": "bridge",    "config": {}  }'

config is a map of string network settings passed to the compute system as is. The response is 201 with the network object. A name already in use returns 409. If the compute system refuses the network (for example, because the deployment does not allow tenant networks), the request fails and nothing is saved.

List, view and delete networks

• GET /api/v1/compute/networks returns a page of networks (limit, offset). It lists networks created in your tenant through Lahijan; the operator's shared network is not included.
• GET /api/v1/compute/networks/{networkId} returns one network with its type, config and any aclNames and forwardNames.
• DELETE /api/v1/compute/networks/{networkId} deletes it and returns 204.

There is no update call. Detach every instance from a network before you delete it.

Attach an instance to a network

Add a nic device that names the network. In the dashboard, open the instance's Config tab and, under Instance devices, add a device:

• Name: for example eth1

• Type: nic

• Properties:

  inicopy

  network=backend

With the API, send the device in a PATCH. The devices map replaces all of the instance's own devices, so include any you already have:

curl -X PATCH https://lahijan.example.com/api/v1/compute/instances/$INSTANCE_ID \  -H "Authorization: Bearer $LAHIJAN_TOKEN" \  -H "X-Tenant-Id: $TENANT_ID" \  -H "Content-Type: application/json" \  -d '{ "devices": { "eth1": { "type": "nic", "network": "backend" } } }'

NIC devices must point at a managed network with network=.... Other NIC kinds are refused by the server.

To give several instances the same network setup, put the nic device in a profile.

Public addresses

To reach an instance from the internet on a fixed public address, allocate a floating IP and attach it to the instance.

Permissions