[LibreQoS] network.json and ShapedDevices.csv for Multi-Wan SOHO

[Dave Taht]

On Mon, Mar 20, 2023 at 7:15 PM Dave Taht <dave.taht at gmail.com> wrote:
>
> The fastest way to get support is via the #libreqos:matrix.org
> chatroom, although we also do github bugs and good ole fashioned
> email.
>
> On Mon, Mar 20, 2023 at 7:03 PM Mike Loiterman via LibreQoS
> <libreqos at lists.bufferbloat.net> wrote:
> >
> > Hello, all.
> >
> > Really new to LibreQoS, but so far it’s been working really, really well with my pfSense setup.
>
> So wonderful to see you trying it. Are you using v1.3, v1.4-rc2 or the
> heimdall (v1.5) branch?
>
> > I’d like some input on how best to setup my network.json and ShapedDevices.csv given the following parameters:

In re-reading what I just wrote, I misunderstood the topology I think.

> > 1. I have three very asymmetric WANs that connect to my pfSense Edge router/firewall:
> >         a. WAN_1 - 1200/35
> >         b. WAN_2 - 25/2
> >         c. WAN_3 - 50/4
> > 2. pfSense connects to my Layer 3 Aruba 2920 switch via a single 10gig transit link so:
> >         a. pfSense: 172.16.0.1/30
> >         b. 2920: 172.16.0.2/30
> > 3. LibreQoS sits between the pfSense and the 2920 with 10gig layer2 bridge between 172.16.0.1 and 172.16.0.2 setup on an intel X710.
> > 4. Behind pfSense I have the following VLAN layout:
> >         a. VLAN_1 @ 192.168.1.0/24 via WAN_1
> >         b. VLAN_2 @ 192.168.2.0/24 via WAN_1
> >         c. VLAN_3 @ 192.168.3.0/24 via WAN_2
> >         d. VLAN_4 @ 192.168.4.0/24 via WAN_3
> >
> > So, given these parameters, can anyone provide some general tips on how to structure the two files?
>
> It doesn't sound to me as though you need much of a network.json. A
> flat topology will suffice, unless you need to model multiple devices
> further down the chain from VLAN_X.
>
> As for the upstream though, load balancing that is kind of painful and
> more the domain of pfsense than libre.
>
> Additionally, shaping those upstreams properly is problematic, I would
> probably use pfsenses's native fq_codel implementation on each of
> those WAN interfaces to install an upper limit for egress but on
> ingress... hmm... and in either case something needs to attempt to
> direct the "right" amount of traffic through those interfaces.
>
>
> >
> > ------------------------------
> > Mike Loiterman
> > Cell: 630-302-4944
> > Email: mike at ascendency.net
> >
> > _______________________________________________
> > LibreQoS mailing list
> > LibreQoS at lists.bufferbloat.net
> > https://lists.bufferbloat.net/listinfo/libreqos
>
>
>
> --
> Come Heckle Mar 6-9 at: https://www.understandinglatency.com/
> Dave Täht CEO, TekLibre, LLC



-- 
Come Heckle Mar 6-9 at: https://www.understandinglatency.com/
Dave Täht CEO, TekLibre, LLC