[FROG] how to control the metric of routes created by frr?

[Anton Gubar'kov]

Dear fellows, 

I have trouble finding the particular way of increasing the metric of
routes created by frr as a result of the running OSPF.

My network is built as follows:


The virtual backbone is zerotier 10.243/16. Edge routers are connected
to it. Also my laptop is connected to the zerotier backbone (via LTE
modem) and runs FRR. Routers run Mikrotik RouterOS and I use OSPF to
propagate  the routes to 192.168 networks to my Laptop1.   



10.243/16: router1 -------- router2 ------- Laptop1
            |                  |
            |                  |
            |                  |
            |                  |
192.168.201/24             192.168.210/24


Everything works fine until I connect my laptop to one of the 192.168
networks via Wifi.

10.243/16: router1 -------- router2 ------- Laptop1
            |                  |              |
            |                  |              |
            |                  |              |
            |                  |              | 
192.168.201/24             192.168.210/24-----+

By default the metric of the route of the connected Wifi adapter is
600, while the routes that are created by frr have a metric of 20.

anton at PF16W6Y2 ~ $ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use
Iface
default         router          0.0.0.0         UG    600    0        0
wlan0
10.0.85.2       0.0.0.0         255.255.255.255 UH    0      0        0
outline-tun0
10.243.0.0      0.0.0.0         255.255.0.0     U     0      0        0
ztzlge64hc
192.168.122.0   0.0.0.0         255.255.255.0   U     0      0        0
virbr0
192.168.201.0   PERV1CAP1.vpn   255.255.255.0   UG    20     0        0
ztzlge64hc
192.168.210.0   0.0.0.0         255.255.255.0   U     600    0        0
wlan0
192.168.210.0   SOV1.vpn        255.255.255.0   UG    20     0        0
ztzlge64hc


So when I start to pour a lot of data from my laptop to one of the
directly connected servers at 192.168.210/24, the traffic goes to
zerotier interface first, then to router2 and decrypted there and only
after that reaches the server. The CPU util of the router2 jumps to
100% and the speed goes down. 

Is there a way to increase the metric of routes via ztzlge64hc
interface that are created by frr? 
Zerotier client doesn't offer to configure the metric of its interface
and sets it to 0 always. There is a 5-year old feature request to them
for this...

The frr config I have is trivial:

!
frr version 10.1.1-gentoo
frr defaults traditional
hostname PF16W6Y2
log file /var/log/frr/ospfd.log
log syslog informational
no ipv6 forwarding
service integrated-vtysh-config
!
interface ztzlge64hc
 ip ospf area 10.243.0.0 10.243.169.159
 ip ospf cost 1500
exit
!
router ospf
 ospf router-id 10.243.169.159
 log-adjacency-changes
exit
!
end
 

As you can see I tried to increase the cost of the ztzlge64hc interface
and it is correctly reflected in the OSPF routes:

PF16W6Y2# show ip ospf route
============ OSPF network routing table ============
N    10.243.0.0/16         [1500] area: 10.243.0.0
                           directly attached to ztzlge64hc
N    192.168.201.0/24      [1501] area: 10.243.0.0
                           via 10.243.130.110, ztzlge64hc
N    192.168.210.0/24      [2500] area: 10.243.0.0
                           via 10.243.86.42, ztzlge64hc

============ OSPF router routing table =============
R    192.168.201.2         [1500] area: 10.243.0.0, ASBR
                           via 10.243.130.110, ztzlge64hc

============ OSPF external routing table ===========

It doesn't affect the kernel route metrics however...

Thanks for pointing me to the right direction.