Dive into the core networking components of MAAS
Errors or typos? Topics missing? Hard to read? Let us know!
In this chapter, we'll cover everything from subnets to proxies:
Need a TCP/IP refresher? Check out our theory primer first.
Taming MAAS network essentials
Navigate, manage, and optimize your network elements effortlessly, whether you're a UI enthusiast or a CLI aficionado.
Connecting MAAS networks via the UI
Kicking off your networking journey with MAAS? A great first step is enabling network discovery. This feature scans your environment to identify all connected devices, giving you a comprehensive view of your network landscape. It's like a radar system for your digital world, and setting it up is straightforward. Follow our how-to guide to get started.
For MAAS 3.4:
Select Networking > Network discovery in the left navigation panel.
Select the Configuration tab at the top of the Network discovery panel.
In the Network discovery drop-down, select "Enabled".
Select Save to register your changes.
For earlier MAAS versions:
Select Canonical MAAS > Configuration at the top left of the screen.
In the Network discovery drop-down, choose "Enabled" or "Disabled".
Subnet toggling offers you the flexibility to either manage a subnet manually or let MAAS handle it for you. By toggling subnet management, you can decide when to assign IP addresses or set routes for specific subnets. This feature acts like a switch, giving you command over how hands-on or automated you want your network setup to be.
For MAAS 3.4:
Select Subnets from the left navigation panel.
Select the subnet you wish to change by clicking on its address.
Select Edit in the upper right of the Subnet summary panel.
Select Managed allocation to toggle between enabled and disabled.
Select Save to register your changes.
You can (re)enable subnet management at any time by checking Managed allocation.
For earlier MAAS versions:
Select Subnets.
Select the subnet in question
Select Edit.
Select Managed allocation to toggle between enabled and disabled.
Select Save summary.
For MAAS 3.4:
To access the main networking view, select Networking > Subnets. This main view can also be filtered through the use of the 'Filters' drop-down.
For earlier MAAS versions:
To access the main networking view, select Subnets. This main view can also be filtered via the Group by drop-down.
Displaying the details about a subnet with the UI couldn't be simpler. Just select a subnet to view its details.
A little below the Subnet summary, you'll find the utilisation figures for that subnet.
A little further down on the Subnet summary you can see -- and manage -- static routes, with the provided buttons.
For MAAS 3.4:
Select Networking > Subnets.
Select the Subnet you want to change by clicking on its IP address.
In the Subnet summary pane, scroll down to Add static route and select it.
Enter a Gateway IP address.
Select a Destination subnet from the drop-down.
Enter a routing Metric value, if desired.
Select Save to register your changes.
For earlier MAAS versions:
Click the 'Add static route' button to reveal the edit pane.
Enter a Gateway IP address.
Select a destination subnet from the 'Destination' drop-down list.
Edit the routing metric value if needed.
Click 'Add' to activate the route.
Routes can be edited and removed using the icons to the right of each entry.
Near the bottom of the Subnet summary, you'll be able to manage reserved IP ranges.
At the very bottom of the Subnet summary, you can track DHCP snippets and used IP addresses.
To configure a bridge with the MAAS UI:
Select Machines.
Select the machine you want to bridge.
Select Network.
Checkbox the network where you want to create the bridge.
Select Create bridge.
Fill in the bridge details in the form which appears.
Optionally include the bridge in the spanning tree protocol by selecting the slider under Advanced options.
Optionally set the Forward delay in milliseconds.
Select Save interface to register your changes.
You can then deploy machines using this bridge.
You can create an "Open switch" bridge if desired, and MAAS will create the netplan model for you.
Connecting MAAS networks with the CLI
If you prefer the CLI, the same functionality is available, though it takes a slightly different form:
Enable network discovery with the CLI
To enable network discovery, enter the following at the command line:
maas $PROFILE maas set-config name=network_discovery value="enabled"
If successful, you should receive output similar to:
Success.
Machine-readable output follows:
OK
Network discovery can be disabled or re-enabled at any time with this CLI command.
To enable or disable subnet management with the MAAS CLI, enter this command:
maas $PROFILE subnet update $SUBNET_CIDR managed=false|true
For example, to disable subnet management:
maas $PROFILE subnet update 192.168.1.0/24 managed=false
You can use the subnets ID in place of the CIDR address.
You'll need the "fabric ID" to manipulate some networking parameters in the CLI. To determine a fabric ID based on a subnet address:
FABRIC_ID=$(maas $PROFILE subnet read $SUBNET_CIDR \
| grep fabric | cut -d ' ' -f 10 | cut -d '"' -f 2)
This may come in handy when you need a fabric ID for other CLI calls.
To set the default gateway for a subnet, just use this command:
maas $PROFILE subnet update $SUBNET_CIDR gateway_ip=$MY_GATEWAY
You can set the DNS server for a subnet very easily, with the CLI, like this:
maas $PROFILE subnet update $SUBNET_CIDR dns_servers=$MY_NAME SERVER
From time to time, you may want to view the list of available subnets. Do that with the following command:
maas admin subnets read | \
jq -r '(["FABRIC", "VLAN", "DHCP", "SUBNET"]
| (., map(length*"-"))),
(.[] | [.vlan.fabric, .vlan.name, .vlan.dhcp_on, .cidr])
| @tsv' \
| column -t
Via the CLI, you can view the details of an individual subnet with the command:
maas $PROFILE subnet read $SUBNET_ID \
| jq -r '(["NAME","CIDR","GATEWAY","DNS","DISCOVERY","FABRIC","VLAN"]
| (., map(length*"-"))), ([.name,.cidr,.gateway_ip // "-", .allow_dns,.active_discovery,.vlan.name,.vlan.fabric]) | @tsv' | column -t
Look up the subnet ID like this:
maas $PROFILE subnets read \
| jq -r '(["NAME", "SUBNET_ID"]
| (., map(length*"-"))), (.[] | [.name, .id]) | @tsv' \
| column -t | grep $SUBNET_NAME
For example, using the "admin" profile with a subnet name containing "192.168.123," find the subnet ID with this command:
maas admin subnets read \
| jq -r '(["NAME", "SUBNET_ID"]
| (., map(length*"-"))), (.[] | [.name, .id]) | @tsv' \
| column -t | grep 192.168.123
With the CLI, it's simple create a static route between two subnets:
maas admin static-routes create source=$SOURCE_SUBNET destination=$DEST_SUBNET \
gateway_ip=$GATEWAY_IP
You can use the MAAS CLI/API to configure a bridge via the following procedure:
Select the interface on which you wish to configure the bridge. This example uses the boot interface, since the boot interface must be connected to a MAAS controlled network -- but any interface is allowed:
INTERFACE_ID=$(maas $PROFILE machine read $SYSTEM_ID | jq .boot_interface.id)
Create the bridge:
BRIDGE_ID=$(maas $PROFILE interfaces create-bridge $SYSTEM_ID name=br0 parent=$INTERFACE_ID | jq .id)
Select the subnet where you want the bridge (this should be a MAAS controlled subnet):
SUBNET_ID=$(maas $PROFILE subnets read | jq -r '.[] | select(.cidr == "10.0.0.0/24" and .managed == true).id')
Connect the bridge to the subnet:
maas $PROFILE interface link-subnet $SYSTEM_ID $BRIDGE_ID subnet=$SUBNET_ID mode="STATIC" ip_address="10.0.0.101"
Building a bridge with netplan is independent of MAAS UI or CLI. You can configure a bridge like this:
Open your netplan configuration file. This should be in /etc/netplan. It could be called 50-cloud-init.yaml, netplan.yaml, or something else.
Modify the file to add a bridge, using the following example as a guide:
network:
bridges:
br0:
addresses:
- 10.0.0.101/24
gateway4: 10.0.0.1
interfaces:
- enp1s0
mac address: 52:54:00:39:9d:f9
mtu: 1500
name servers:
addresses:
- 10.0.0.2
search:
- maas
parameters:
forward-delay: 15
stp: false
Ethernet's:
enp1s0:
match:
mac address: 52:54:00:39:9d:f9
mtu: 1500
set-name: enp1s0
enp2s0:
match:
mac address: 52:54:00:df:87:ac
mtu: 1500
set-name: enp2s0
enp3s0:
match:
mac address: 52:54:00:a7:ac:46
mtu: 1500
set-name: enp3s0
version: 2
netplan apply.You have a lot of latitude to change machine interfaces with MAAS. In this section, you can learn the ins and outs of machine interfaces:
3.4 UI
To edit a machine interface with the 3.4 UI:
Select Machines.
Select the machine in question by clicking on its name.
Select Network.
Checkbox the interface in question.
Click the Actions drop-down at the right end of row for that interface.
Select Edit physical.
Change any of the desired parameters in the form which appears.
Select an IP mode.
Select Save interface to register your changes.
3.3 and below UI
To edit a machine interface with the UI for MAAS 3.3 and below:
From a machine's "Interfaces" page, click the menu icon for the interface to be edited and select Edit Physical from the resulting menu.
Select an IP mode from the drop-down menu.
Press Save to apply the changes.
CLI, all versions
To edit the IP assignment mode of a network interface with the CLI, perform the following steps:
maas $PROFILE node read $SYSTEM_ID
maas $PROFILE interface unlink-subnet $SYSTEM_ID $INTERFACE_ID id=$SUBNET_LINK_ID
maas $PROFILE interface link-subnet $SYSTEM_ID $INTERFACE_ID mode=$IP_MODE subnet=$SUBNET_CIDR [$OPTIONS]
See the glossary for the definitions of reserved range types.
With the UI:
Select more than one interface.
Select Create bond; the bond configuration pane will appear.
Rename the bond, if desired.
Select a bond mode:
balance-rr: Transmit packets in sequential order from the first available follower through to the last. This mode provides load balancing and fault tolerance.
active-backup: Only one follower in the bond is active. A different follower becomes active if, and only if, the active follower fails. The bond's MAC address is externally visible on only one port (network adaptor) to avoid confusing the switch.
balance-xor: Transmit based on the selected transmit hash policy. The default policy is simple, which means that an XOR operation selects packages. This XOR compares the source MAC address and the resultant XOR between the destination MAC address, the packet type identifier, and the modulo follower count.
broadcast: Transmit everything on all follower interfaces. This mode provides fault tolerance.
802.3ad: Creates aggregation groups that share the same speed and duplex settings. This mode utilises all followers in the active aggregation, following the IEEE 802.3ad specification.
balance-tlb: Adaptive transmit load balancing, channel bonding that does not require any special switch support.
balance-alb: Adaptive load balancing, includes balance-tlb plus receive load balancing (rlb) for IPV4 traffic. This mode does not require any special switch support. ARP negotiation achieves load balancing in this case.
Assign a MAC address to the aggregate device.
Attach one or more Tags, if desired.
Select the Primary device.
Select Save to register your changes.
The MAC address defaults to the MAC address of the primary interface.
Create a bond with the CLI:
To create a bond, execute the following command:
maas $PROFILE interfaces create-bond $SYSTEM_ID name=$BOND_NAME \
parents=$IFACE1_ID mac_address=$MAC_ADDR \
parents=$IFACE2_ID bond_mode=$BOND_MODE \
bond_updelay=$BOND_UP bond_downdelay=$BOND_DOWN mtu=$MTU
Note that:
The parents parameters define which interfaces form the aggregate interface.
The bond_updelay and bond_downdelay parameters specify the number of milliseconds to wait before either enabling or disabling a follower after a failure has been detected.
There are a wide range of bond parameters you can choose when creating a bond:
| Parameter | Type and description |
|---|---|
mac_address |
Optional string. MAC address of the interface. |
tags |
Optional string. Tags for the interface. |
vlan |
Optional string. VLAN the interface is connected to. If not provided then the interface is considered disconnected. |
parents |
Required integer. Parent interface ids that make this bond. |
bond_miimon |
Optional integer. The link monitoring frequency in milliseconds. (Default: 100). |
bond_downdelay |
Optional integer. Specifies the time, in milliseconds, to wait before disabling a follower after a link failure has been detected. |
bond_updelay |
Optional integer. Specifies the time, in milliseconds, to wait before enabling a follower after a link recovery has been detected. |
bond_lacp_rate |
Optional string. Option specifying the rate at which to ask the link partner to transmit LACPDU packets in 802.3ad mode. Available options are fast or slow. (Default: slow). |
bond_xmit_hash_policy |
Optional string. The transmit hash policy to use for follower selection in balance-xor, 802.3ad, and tlb modes. Possible values are: layer2, layer2+3, layer3+4, encap2+3, encap3+4. (Default: layer2) |
bond_num_grat_arp |
Optional integer. The number of peer notifications (IPv4 ARP or IPv6 Neighbour Advertisements) to be issued after a failover. (Default: 1) |
mtu |
Optional integer. Maximum transmission unit. |
accept_ra |
Optional Boolean. Accept router advertisements. (IPv6 only) |
autoconf |
Optional Boolean. Perform stateless autoconfiguration. (IPv6 only) |
bond_mode |
Optional string. The operating mode of the bond. (Default: active-backup). |
| Mode | Behaviour |
|---|---|
balance-rr: |
Transmit packets in sequential order from the first available follower through the last. This mode provides load balancing and fault tolerance. |
active-backup |
Only one follower in the bond is active. A different follower becomes active if, and only if, the active follower fails. The bond's MAC address is externally visible on only one port (network adaptor) to avoid confusing the switch. |
balance-xor |
Transmit based on the selected transmit hash policy. The default policy is a simple [(source MAC address XOR'd with destination MAC address XOR packet type ID) modulo follower count]. |
broadcast |
Transmits everything on all follower interfaces. This mode provides fault tolerance. |
802.3ad |
IEEE 802.3ad dynamic link aggregation. Creates aggregation groups that share the same speed and duplex settings. Uses all followers in the active aggregator according to the 802.3ad specification. |
balance-tlb |
Adaptive transmit load balancing: channel bonding that does not require any special switch support. |
balance-alb |
Adaptive load balancing: includes balance-tlb plus receive load balancing (rlb) for IPV4 traffic, and does not require any special switch support. The receive load balancing is achieved by ARP negotiation. |
With the UI:
To create a bridge interface:
Select Machines.
Select a machine.
Select Network.
Select an interface by choosing its checkbox.
Select Create bridge.
Optionally enter a unique Bridge name.
Choose a Bridge type.
Enter a valid MAC address.
Choose a Fabric.
Choose a VLAN.
Optionally choose a Subnet.
Optionally enter Tags.
Optionally turn on STP.
Register your new bridge by selecting Save interface.
With the CLI:
To create a bridge interface with the CLI:
maas $PROFILE interfaces create-bridge $SYSTEM_ID name=$BRIDGE_NAME \
parent=$IFACE_ID
Use parent to define the primary interface used for the bridge:
maas admin interfaces create-bridge 4efwb4 name=bridged0 parent=4
The following parameters may be applied when creating a bridge:
name: Optional string. Name of the interface.
mac_address: Optional string. MAC address of the interface.
tags: Optional string. Tags for the interface.
vlan: Optional string. VLAN the interface is connected to.
parent: Optional integer. Parent interface id for this bridge interface.
bridge_type: Optional string. The type of bridge to create. Possible values are: standard, ovs.
bridge_stp: Optional Boolean. Turn spanning tree protocol on or off. (Default: False).
bridge_fd: Optional integer. Set bridge forward delay to time seconds. (Default: 15).
mtu: Optional integer. Maximum transmission unit.
accept_ra: Optional Boolean. Accept router advertisements. (IPv6 only)
autoconf: Optional Boolean. Perform stateless autoconfiguration. (IPv6 only)
Interfaces can only be deleted via the CLI. The "delete" command can be used to delete a bridge interface, a bond interface or a physical interface:
maas $PROFILE interface delete $SYSTEM_ID $IFACE_ID
For example:
maas admin interface delete 4efwb4 15
The following is output after the successful deletion of an interface:
Success.
Machine-readable output follows:
Note that while the label is presented, there is no machine-readable output expected after the successful execution of the delete command.
Assigning an interface to a fabric
You can only assign an interface to a fabric with the MAAS CLI. This task is made easier with the aid of the jq utility. It filters the maas command (JSON formatted) output and prints it in the desired way, which allows you to view and compare data quickly. Go ahead and install it:
sudo apt install jq
In summary, MAAS assigns an interface to a fabric by assigning it to a VLAN. First, we need to gather various bits of data.
List some information on all machines:
maas $PROFILE machines read | jq ".[] | \
{hostname:.hostname, system_id: .system_id, status:.status}" --compact-output
Example output:
{"hostname":"machine1","system_id":"dfgnnd","status":4}
{"hostname":"machine2","system_id":"bkaf6e","status":6}
{"hostname":"machine4","system_id":"63wqky","status":6}
{"hostname":"machine3","system_id":"qwkmar","status":4}
You can only edit an interface when the corresponding machine has a status of 'Ready'. This state is numerically denoted by the integer '4'.
List some information for all interfaces on the machine in question (identified by its system id 'dfgnnd'):
maas $PROFILE interfaces read dfgnnd | jq ".[] | \
{id:.id, name:.name, mac:.mac_address, vid:.vlan.vid, fabric:.vlan.fabric}" --compact-output
Example output:
{"id":8,"name":"eth0","mac":"52:54:00:01:01:01","vid":0,"fabric":"fabric-1"}
{"id":9,"name":"eth1","mac":"52:54:00:01:01:02","vid":null,"fabric":null}
List some information for all fabrics:
maas $PROFILE fabrics read | jq ".[] | \
{name:.name, vlans:.vlans[] | {id:.id, vid:.vid}}" --compact-output
Example output:
{"name":"fabric-0","vlans":{"id":5001,"vid":0}}
{"name":"fabric-1","vlans":{"id":5002,"vid":0}}
{"name":"fabric-2","vlans":{"id":5003,"vid":0}}
This example will show how to move interface '8' (on machine 'dfgnnd') from 'fabric-1' to 'fabric-0'. Based on the gathered information, this will consist of changing the interface's VLAN from '5002' to '5001':
maas $PROFILE interface update dfgnnd 8 vlan=5001 >/dev/null
Verify the operation by relisting information for the machine's interface:
maas $PROFILE interfaces read dfgnnd | jq ".[] | \
{id:.id, name:.name, mac:.mac_address, vid:.vlan.vid, fabric:.vlan.fabric}" --compact-output
The output shows that the interface is now on fabric-0:
{"id":8,"name":"eth0","mac":"52:54:00:01:01:01","vid":0,"fabric":"fabric-0"}
{"id":9,"name":"eth1","mac":"52:54:00:01:01:02","vid":null,"fabric":null}
Interface identifiers can only be discovered via the MAAS CLI. The MAAS CLI uses a numeric interface identifier for many interface operations. Use the following command to retrieve the identifier(s):
maas $PROFILE interfaces read $SYSTEM_ID
Look for either id or the number at the end of an interface's resource URI, such as 15 in the following example output:
"id": 15,
"mac_address": "52:54:00:55:06:40",
...
"name": "ens9",
...
"resource_uri": "/MAAS/api/2.0/nodes/4efwb4/interfaces/15/"
VLAN interfaces can only be created via the MAAS CLI. To create a VLAN interface, use the following syntax:
maas $PROFILE vlans create $FABRIC_ID name=$NAME vid=$VLAN_ID
For example, the following command creates a VLAN called 'Storage network:
maas admin vlans create 0 name="Storage network" vid=100
The above command generates the following output:
Success.
Machine-readable output follows:
{
"vid": 100,
"mtu": 1500,
"dhcp_on": false,
"external_dhcp": null,
"relay_vlan": null,
"name": "Storage network",
"space": "undefined",
"fabric": "fabric-0",
"id": 5004,
"primary_rack": null,
"fabric_id": 0,
"secondary_rack": null,
"resource_uri": "/MAAS/api/2.0/vlans/5004/"
}
Be aware that the $VLAN_ID parameter does not indicate a VLAN ID that corresponds to the VLAN tag. You must first create the VLAN and then associate it with the interface:
maas $PROFILE interfaces create-vlan $SYSTEM_ID vlan=$OUTPUT_VLAN_ID \
parent=$IFACE_ID
OUTPUT_VLAN_ID corresponds to the id value output when MAAS created the VLAN.
The following example contains values that correspond to the output above:
maas admin interfaces create-vlan 4efwb4 vlan=5004 parent=4
The above command generates the following output:
Success.
Machine-readable output follows:
{
"tags": [],
"type": "vlan",
"enabled": true,
"system_id": "4efwb4",
"id": 21,
"children": [],
"mac_address": "52:54:00:eb:f2:29",
"params": {},
"vlan": {
"vid": 100,
"mtu": 1500,
"dhcp_on": false,
"external_dhcp": null,
"relay_vlan": null,
"id": 5004,
"secondary_rack": null,
"fabric_id": 0,
"space": "undefined",
"fabric": "fabric-0",
"name": "Storage network",
"primary_rack": null,
"resource_uri": "/MAAS/api/2.0/vlans/5004/"
},
"parents": [
"ens3"
],
"effective_mtu": 1500,
"links": [
{
"id": 55,
"mode": "link_up"
}
],
"discovered": null,
"name": "ens3.100",
"resource_uri": "/MAAS/api/2.0/nodes/4efwb4/interfaces/21/"
}
VLAN interfaces can only be deleted via the MAAS CLI. The following command outlines the syntax required to delete a VLAN interface from the command line:
maas $PROFILE vlan delete $FABRIC__ID $VLAN_ID
Using the values from previous examples, you executed this step as follows:
maas admin vlan delete 0 100
MAAS lets you set up a proxy for managed machines to access web resources. Your choices:
Toggle these options and control proxy use by subnet. This section will show you how.
Set up a proxy with the UI:
In the web UI, visit the 'Settings' page and select the 'Network services' tab.
Modify the 'Proxy' section at the top, as desired:
To enable the internal proxy, ensure that the checkbox adjacent to 'MAAS Built-in' is selected. This internal proxy is the default configuration.
To enable an external proxy, activate the 'External' checkbox and use the new field that is displayed to define the proxy's URL (and port if necessary).
An upstream cache peer can be defined by enabling the 'Peer' checkbox and entering the external proxy URL into the field. With this enabled, machines will be configured to use the MAAS built-in proxy to download cached APT packages.
To prevent MAAS machines from using a proxy, enable the 'Don't use a proxy' checkbox.
Set up a proxy with the CLI
Enabling and disabling proxying, in general, is done via a Boolean option ('true' or 'false'). The following command will disable proxying completely:
maas $PROFILE maas set-config name=enable_http_proxy value=false
To set an external proxy, ensure proxying is enabled (see above) and then define it:
maas $PROFILE maas set-config name=http_proxy value=$EXTERNAL_PROXY
For example,
maas $PROFILE maas set-config name=enable_http_proxy value=true
maas $PROFILE maas set-config name=http_proxy value=http://squid.example.com:3128/
Enabling and disabling proxying per subnet is done via a Boolean option ('true' or 'false'). Here is how you can disable proxying on a per-subnet basis:
maas $PROFILE subnet update $SUBNET_CIDR allow_proxy=false
For example,
maas $PROFILE subnet update 192.168.0.0/22 allow_proxy=false
NOTE that the proxy service will still be running.
MAAS provides managed NTP services (with Chrony^) for all region and rack controllers. This arrangement allows MAAS to both keep its controllers synchronised, and keep deployed machines synchronised as well. You can configure NTP on the 'Network services' tab of the 'Settings' page.
The region controller configures the NTP service to keep its time synchronised from one or more external sources. By default, the MAAS region controller uses ntp.ubuntu.com. Rack controllers also configure the NTP service, synchronising their time with the region controllers. Rack controllers also configure DHCP with the correct NTP information, so that the DHCP servers can manage the NTP clients for the rack. Any machine on the network that obtains a DHCP lease from MAAS will benefit from NTP support.
Setting an external NTP server
External sites, such as an existing NTP infrastructure, can be used directly as a time source for both rack controllers and machines.
Set up an NTP server with the UI:
You can specify an external site by choosing the NTP server(s) and selecting the 'External Only' option. The region controller always uses an external site.
On the 'Settings' page, select the 'Network services' tab and scroll down to the 'NTP' section.
Enter the address of the desired NTP server.
With the CLI:
You can specify an external NTP server with two successive commands:
maas $PROFILE maas set-config name=ntp_servers value=$NTP_IP_ADDRESS
followed by:
maas admin maas set-config name=ntp_external_only value=true