Management VRF
Management VRF is a subset of Virtual Routing and Forwarding - VRF (virtual routing tables and forwarding) and provides a separation between the out-of-band management network and the in-band data plane network. For VRFs, the main routing table is the default table for the data plane switch ports. With management VRF, the switch uses a second table, mgmt, for routing through the Ethernet ports of the switch. The mgmt name is special cased to identify the management VRF from a data plane VRF.
Cumulus Linux only supports eth0 (or eth1, depending on the switch platform) for out-of-band management. The Ethernet ports are software-only ports that are not hardware accelerated by switchd
. VLAN subinterfaces, bonds, bridges, and the front panel switch ports are not supported as OOB management interfaces.
In band management of Cumulus Linux is possible using loopbacks and SVIs (switch virtual interfaces).
Cumulus Linux enables Management VRF by default. IPv4 and IPv6 networking applications (for example, Ansible, Chef, and apt-get
) run by an administrator communicate out the management network by default. This default context does not impact services run through systemd
and the systemctl
command, and does not impact commands examining the state of the switch, such as the ip
command to list links, neighbors, or routes.
The management VRF configurations in this section contain a localhost loopback IPv4 address of 127.0.0.1/8 and IPv6 address of ::1/128. Management VRF must have an IPv6 address as well as an IPv4 address to work correctly. Adding the loopback address to the layer 3 domain of the management VRF prevents issues with applications that expect the loopback IP address to exist in the VRF, such as NTP.
Bring Up the Management VRF
If you take down the management VRF using ifdown
, to bring it back up you need to do one of two things:
- Run the
ifup --with-depends mgmt
command - Run
ifreload -a
command
The following command example brings down the management VRF, then brings it back up with the ifup --with-depends mgmt
command:
cumulus@switch:~$ sudo ifdown mgmt
cumulus@switch:~$ sudo ifup --with-depends mgmt
Running ifreload -a
disconnects the session for any interface configured as auto.
Run Services within the Management VRF
Most default services in Cumulus Linux are VRF aware. If you want to run a service within the management VRF instead of the default VRF, run the following commands:
If the service is running, stop the service:
cumulus@switch:~$ sudo systemctl stop <service>.service
Disable the service from starting automatically in the default VRF:
cumulus@switch:~$ sudo systemctl disable <service>.service
Start the service in the management VRF:
cumulus@switch:~$ sudo systemctl start <service>@mgmt.service
Enable the service in the management VRF so that it starts when the switch boots:
cumulus@switch:~$ sudo systemctl enable <service>@mgmt.service
Verify that the service is running in the management VRF with the
ps aux | grep <service>
command.
Run the following command to show the process IDs associated with the management VRF:
cumulus@switch:~$ ip vrf pids mgmt
2559 login
2753 bash
2045 dhclient
5421 sshd
5462 sshd
5463 bash
37691 sshd
37732 sshd
37735 bash
55679 sshd
55720 sshd
55721 bash
55993 ip
3834 ntpd
2023 python3
2563 netqd
1855 login
2770 bash
Run the following command to show the VRF association of the specified process:
cumulus@switch:~$ ip vrf identify 2045
mgmt
Run ip vrf help
for additional ip vrf
commands.
Enable Polling with snmpd in a Management VRF
When you enable snmpd
to run in the management VRF, you need to specify that VRF so that snmpd
listens on eth0 in the management VRF; you can also configure snmpd
to listen on other ports. In Cumulus Linux, SNMP configuration is VRF aware so snmpd
can bind to multiple IP addresses each configured with a particular VRF (routing table). The snmpd
daemon responds to polling requests on the interfaces of the VRF on which the request comes in. For information about configuring SNMP version 1, 2c, and 3 Traps and (v3) Inform messages, refer to Simple Network Management Protocol - SNMP.
The message Duplicate IPv4 address detected, some interfaces may not be visible in IP-MIB
displays after starting snmpd
in the management VRF. This is because the IP-MIB assumes that you cannot use the same IP address twice on the same device; the IP-MIB is not VRF aware. This message is a warning that the SNMP IP-MIB detects overlapping IP addresses on the system; it does not indicate a problem and does not impact the operation of the switch.
ping or traceroute on the Management VRF
By default, when you issue a ping
or traceroute
, the packet goes to the data plane network (the main routing table). To use ping
or traceroute
on the management network, use ping -I mgmt
or traceroute -i mgmt
. To select a source address within the management VRF, use the -s
flag for traceroute
.
cumulus@switch:~$ ping -I mgmt <destination-ip>
cumulus@switch:~$ sudo traceroute -i mgmt -s <source-ip> <destination-ip>
For additional information on using ping
and traceroute
, see Network Troubleshooting.
Run Services as a Non-root User
To run services in the management VRF as a non-root user, you need to create a custom service based on the original service file. The following example commands configure the SSH service to run in the management VRF as a non-root user.
Run the following command to create a custom service file in the
/etc/systemd/system
directory.cumulus@switch:~$ sudo -E systemctl edit --full ssh.service
If a User directive exists under [Service], comment it out.
cumulus@switch:~$ sudo nano /etc/systemd/system/ssh.service ... [Service] #User=username ExecStart=/usr/local/bin/ssh agent -data-dir=/tmp/ssh -bind=192.168.0.11 ...
Modify the ExecStart line to
/usr/bin/ip vrf exec mgmt /sbin/runuser -u USER -- ssh
:... [Service] #User=username ExecStart=/usr/bin/ip vrf exec mgmt /sbin/runuser -u cumulus -- ssh ...
OSPF and BGP
FRR is VRF-aware and sends packets based on the switch port routing table. This includes BGP peering through loopback interfaces. BGP looks up routes in the default table. However, depending on how you redistribute your routes, you can perform the following modification.
Management VRF uses the mgmt table, including local routes. This does not affect route redistribution when you use routing protocols, such as OSPF and BGP.
To redistribute the routes in your network, use the redistribute connected
command under BGP or OSPF. This enables the directly connected network out of eth0 to advertise to its neighbor.
This also creates a route on the neighbor device to the management network through the data plane.
NVIDIA recommends route maps to control advertised networks that you redistribute with the redistribute connected
command.
cumulus@switch:~$ nv set router policy route-map REDISTRIBUTE rule 10 match type ipv4
cumulus@switch:~$ nv set router policy route-map REDISTRIBUTE rule 10 match interface eth0
cumulus@switch:~$ nv set router policy route-map REDISTRIBUTE rule 10 action deny
cumulus@switch:~$ nv set vrf default router bgp address-family ipv4-unicast redistribute connected route-map REDISTRIBUTE
cumulus@switch:~$ nv config apply
cumulus@switch:$ sudo vtysh
...
switch# configure terminal
switch(config)# route-map REDISTRIBUTE-CONNECTED deny 10
switch(config-route-map)# match interface eth0
switch(config)# route-map REDISTRIBUTE-CONNECTED permit 100
switch(config-route-map)# exit
switch(config)# router bgp
switch(config-router)# address-family ipv4 unicast
switch((config-router-af)# redistribute connected route-map REDISTRIBUTE-CONNECTED
switch(config)# end
switch# write memory
switch# exit
The vtysh commands save the configuration in the /etc/frr/frr.conf
file. For example:
...
router bgp 65101
bgp router-id 10.10.10.1
neighbor swp51 interface remote-as external
neighbor swp52 interface remote-as external
!
address-family ipv4 unicast
network 10.1.10.0/24
network 10.10.10.1/32
redistribute connected route-map REDISTRIBUTE-CONNECTED
maximum-paths 64
maximum-paths ibgp 64
exit-address-family
!
route-map REDISTRIBUTE-CONNECTED deny 100
match interface eth0
!
route-map REDISTRIBUTE-CONNECTED permit 1000
...
SSH
To limit SSH to listen on only the management VRF or to a specific IP address on the management VRF, see SSH and VRFs.
If you SSH to the switch through a switch port, SSH works as expected. If you need to SSH from the switch out of a switch port, use the
ip vrf exec default ssh <switch-port-ip-address>
command. For example:cumulus@switch:~$ sudo ip vrf exec default ssh 10.3.3.3
View the Routing Tables
When you use ip route get
to return information about a single route, the command resolves over the mgmt table by default. To show information about the route in the switching silicon, run this command:
cumulus@switch:~$ ip route get <ip-address>
To get the route for any VRF, run the ip route get <ip-address> oif <vrf-name>
command. For example, to show the route for the management VRF, run:
cumulus@switch:~$ ip route get <ip-address> oif mgmt
mgmt Interface Class
ifupdown2
uses interface classes to create a user-defined grouping for interfaces. The special class mgmt is available to separate the management interfaces of the switch from the data interfaces. This allows you to manage the data interfaces by default using ifupdown2
commands. Performing operations on the mgmt interfaces requires specifying the --allow-mgmt
option, which prevents inadvertent outages on the management interfaces. Cumulus Linux by default brings up all interfaces in both the auto (default) class and the mgmt interface class when the switch boots.
You configure the management interface in the /etc/network/interfaces
file. The example below adds the management interface eth0 and the management VRF stanzas to the mgmt interface class:
...
auto lo
iface lo inet loopback
allow-mgmt eth0
iface eth0 inet dhcp
vrf mgmt
allow-mgmt mgmt
iface mgmt
address 127.0.0.1/8
address ::1/128
vrf-table auto
...
When you run ifupdown2
commands against the interfaces in the mgmt class, include --allow=mgmt
with the commands. For example, to see which interfaces are in the mgmt interface class, run:
cumulus@switch:~$ ifquery l --allow=mgmt
eth0
mgmt
To reload the configurations for interfaces in the mgmt class, run:
cumulus@switch:~$ sudo ifreload --allow=mgmt
You can still bring the management interface up and down using ifup eth0
and ifdown eth0
.
Management VRF and DNS
Cumulus Linux supports both DHCP and static DNS entries over management VRF through IP FIB rules, which it adds to direct lookups to the DNS addresses out of the management VRF.
For DNS to use the management VRF, the static DNS entries must reference the management VRF in the /etc/resolv.conf
file. You cannot specify the same DNS server address twice to associate it with different VRFs.
For example, to specify DNS servers and associate some of them with the management VRF, run the following commands:
cumulus@switch:~$ nv set service dns default server 192.0.2.1
cumulus@switch:~$ nv set service dns mgmt server 198.51.100.31
cumulus@switch:~$ nv set service dns mgmt server 203.0.113.13
cumulus@switch:~$ nv config apply
Edit the /etc/resolv.conf
file to add the DNS servers and associate some of them with the management VRF. For example:
cumulus@switch:~$ sudo nano /etc/resolv.conf
nameserver 192.0.2.1
nameserver 198.51.100.31 # vrf mgmt
nameserver 203.0.113.13 # vrf mgmt
Run the ifreload -a
command to load the new configuration:
cumulus@switch:~$ ifreload -a
- Because FIB rules force DNS lookups out of the management interface, this can affect data plane ports if you use overlapping addresses. For example, when the switch learns the DNS server IP address over the management VRF, it creates a FIB rule for that IP address. When DHCP relay has the same IP address, the switch forwards any DHCP discover packet arriving on the front panel port out of the management interface (eth0) even though a route is present out the front-panel port.
- If you do not specify a DNS server and you lose in band connectivity, DNS does not work through the management VRF. Cumulus Linux does not assume all DNS servers are reachable through the management VRF.