IS-IS on FRR - Is Anyone Running It?

Mark Tinka mark.tinka at seacom.mu
Sat Apr 4 15:16:57 UTC 2020


On 3/Apr/20 21:28, Eduardo Schoedler wrote:
> Mark,
>
> Did you tried this:
> https://lists.freebsd.org/pipermail/freebsd-current/2006-December/068011.html
>
> There are some knobs for Freebsd:
> http://nginx.org/en/docs/freebsd_tuning.html

So the problem really isn't FreeBSD itself. IS-IS looks at MTU in order
to work, and if nothing is manually set, it infers the MTU from the
physical interface over which it is enabled.

While the current IS-IS implementation in FRR is buggy to the extent
that it does not assume MTU can be larger than 8,192 bytes, that does
not prevent an operator from telling it what MTU it should use for IIH
messages, provided the physical interface can support it.

Suffice it to say, I already have a number of Sysctl and kernel knobs in
FreeBSD to tune the system for the Anycast services we run on them. I'd
be disinclined to mess about with that as I don't think it has any
bearing on an over-the-top service such as IS-IS.

Quagga runs well (OSFP though, I'll admit), and I'll keep looking for
answers on IS-IS in FRR until it stops making sense.

Mark.



More information about the NANOG mailing list