Configuring Cumulus Linux and FRRouting

Most networking devices¹ use a configuration command line interface (CLI) to interact with the end-user. The CLI usually provides show commands to inspect the state of the device and a configuration mode that allows the user to configure the device.

Cumulus Linux and FRRouting are different. They are implemented as an application/management layer on top of Linux, and Linux shell is used as the initial CLI. To configure them, you have to:

• Use standard Linux tools like ifupdown to configure the interfaces;
• Edit FRRouting configuration files to start routing protocol daemons
• Start FRRouting configuration shell from the Linux CLI.

The Linux interfaces and IP addresses will be configured automatically if you start the BGP labs with the netlab up command. You will have to start the routing protocol daemons in the initial lab exercises if you plan to use Cumulus Linux or FRRouting as the user routers, and you might have to execute show commands on Cumulus Linux or FRRouting acting as the external routers. You’ll practice both in this lab exercise.

Start the Lab

Assuming you already set up your lab infrastructure:

• Change directory to basic/0-frrouting
• Execute netlab up to start a lab with two Cumulus Linux or FRR virtual machines or containers (depending on your lab setup).

• Log into the devices (rtr and x1) with the netlab connect command.

Start the BGP Daemon

Most network devices start routing daemons when you configure them through the configuration CLI or API. FRRouting (the routing daemons used in Cumulus Linux) is different: you have to enable the desired routing daemons in a configuration file and restart the top-level FRRouting process.

You can check the FRR daemons running on your Cumulus Linux devices or FRR running in a Linux virtual machine with the sudo systemctl status frr.service command. It displays the running FRR daemons and the recent FRR logging messages, for example²:

rtr(bash)#sudo systemctl status frr.service
* frr.service - FRRouting
   Loaded: loaded (/lib/systemd/system/frr.service; enabled; vendor preset: enabled)
   Active: active (running) since Wed 2024-06-26 17:09:37 UTC; 2min 27s ago
     Docs: https://frrouting.readthedocs.io/en/latest/setup.html
  Process: 2230 ExecReload=/usr/lib/frr/frrinit.sh reload (code=exited, status=0/SUCCESS)
   Status: "FRR Operational"
    Tasks: 17 (limit: 737)
   Memory: 19.2M
      CPU: 690ms
   CGroup: /system.slice/frr.service
           |-1875 /usr/lib/frr/zebra -d -F datacenter -M cumulus_mlag -M snmp -A 127.0.0.1 -s 90000000
           |-1889 /usr/lib/frr/staticd -d -F datacenter -A 127.0.0.1
           |-2246 /usr/lib/frr/watchfrr -d -F datacenter zebra bgpd staticd
           `-2255 /usr/lib/frr/bgpd -d -F datacenter -M snmp -A 127.0.0.1

Jun 26 17:10:01 cc bgpd[2255]: Configuration Read in Took: 00:00:00
Jun 26 17:10:01 cc staticd[1889]: Configuration Read in Took: 00:00:00
Jun 26 17:10:03 cc bgpd[2255]: %ADJCHANGE: neighbor 172.16.1.4(fc) in vrf default Up

You cannot use the same command in FRR containers as they don’t use systemd. The easiest way to find FRR daemons in FRR containers is to use the ps -ef|grep frr command³. This is the printout you could get when the BGP daemon is already running:

rtr(bash)#ps -ef|grep frr
    1 root      0:00 /sbin/tini -- /usr/lib/frr/docker-start
   26 root      0:00 /usr/lib/frr/watchfrr -d -F datacenter zebra bgpd staticd
   41 frr       0:00 /usr/lib/frr/zebra -d -F datacenter -A 127.0.0.1 -s 90000000
   46 frr       0:00 /usr/lib/frr/bgpd -d -F datacenter -A 127.0.0.1
   53 frr       0:00 /usr/lib/frr/staticd -d -F datacenter -A 127.0.0.1
  342 root      0:00 grep frr

The list of FRRouting daemons you want to enable is stored in the /etc/frr/daemons file. To enable the FRRouting BGP daemon, you have to:

• Add the bgpd=yes line to the /etc/frr/daemons file⁴.
• Restart FRRouting with the sudo systemctl restart frr.service command (see also: using sudo)

You could add the required line to the FRRouting daemons file with any text editor⁷ or use the following trick:

• Use sudo bash to start another Linux shell as the root user
• Use the echo command with output redirection to add a line to the /etc/frr/daemons file.

rtr(bash)#sudo bash
root@rtr:/# echo 'bgpd=yes' >>/etc/frr/daemons
root@rtr:/# exit

After enabling the BGP daemon and restarting FRR, you should see the bgpd process in the ps -ef printout or the BGP daemon mentioned in the sudo systemctl status frr.service printout.

Work with the FRRouting CLI

FRRouting suite includes a virtual shell (vtysh) closely resembling industry-standard CLI⁸. It has to be started from the Linux command line with the vtysh command. The vtysh CLI has to run as the root user unless you change the FRR-related permissions to allow a regular user to use it. The usual command to start the vtysh is thus sudo vtysh (but see also To Sudo Or Not to Sudo).

rtr(bash)#sudo vtysh

Hello, this is FRRouting (version 7.5+cl4.4.0u4).
Copyright 1996-2005 Kunihiro Ishiguro, et al.

rtr#

Once you started vtysh, you can execute show commands to inspect the device state, for example:

x1(bash)#sudo vtysh

Hello, this is FRRouting (version 7.5+cl4.4.0u4).
Copyright 1996-2005 Kunihiro Ishiguro, et al.

x1# show ip bgp
BGP table version is 1, local router ID is 10.0.0.2, vrf id 0
Default local pref 100, local AS 65100
Status codes:  s suppressed, d damped, h history, * valid, > best, = multipath,
               i internal, r RIB-failure, S Stale, R Removed
Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self
Origin codes:  i - IGP, e - EGP, ? - incomplete

   Network          Next Hop            Metric LocPrf Weight Path
*> 10.0.0.2/32      0.0.0.0                  0         32768 i

Displayed  1 routes and 1 total paths

To configure FRRouting daemons, use the configure vtysh command and enter configuration commands similar to those you’d use on Cisco IOS or Arista EOS:

x1(bash)#sudo vtysh

Hello, this is FRRouting (version 7.5+cl4.4.0u4).
Copyright 1996-2005 Kunihiro Ishiguro, et al.

x1# configure
x1(config)# router bgp 65100
x1(config-router)#

To Sudo Or Not to Sudo

The vtysh usually has to run as the root user, so you should start it with the sudo vtysh command. Unfortunately, things are never as simple as they look:

• When using SSH, you log into Cumulus Linux or FRRouting virtual machines as a regular user (user vagrant in netlab-created labs) and have to use the sudo command to start vtysh.
• Cumulus Linux and FRR containers run as the root user, and you connect to them as the root user with the docker exec or netlab connect commands⁹. When working with containers, you can start vtysh without using the sudo command.
• You can execute sudo vtysh as a root user on Cumulus Linux containers but not within an FRR container. The FRR container does not include the sudo command.

Long story short:

• Use sudo vtysh whenever possible to burn it into your muscle memory.
• Use vtysh if you use FRRouting containers as the lab devices.

Using Output Filters

Unlike many other network operating systems, FRR vtysh does not have output filters. You probably don’t need them as you’ll be running FRR on top of a Unix-like operating system that supports pipes, but it might be a bit convoluted to use vtysh in a pipe.

To use the vtysh output in a pipe, you have to execute vtysh and get the results of a show command in a single command:

• You could use sudo vtysh -c 'show command' when you’re in the bash shell of a lab device, for example:

$ sudo vtysh -c 'show ip bgp' | grep 32768
*> 192.168.100.0/24 0.0.0.0                  0         32768 i

• Alternatively, you could use the netlab connect --show command to execute a vtysh show command on a lab device:

$ netlab connect x1 --show ip bgp | grep 32768
Connecting to container clab-originate-x1, executing sudo vtysh -c "show ip bgp"
*> 192.168.100.0/24 0.0.0.0                  0         32768 i

The following table contains a mapping between common network operating system filters and Linux CLI commands:

NOS filter	Linux CLI command
include	grep
exclude	grep -v
begin	grep -A 1000010
end	grep -B 10000
section	no equivalent