Re: [lime] QoS in Libremesh

Borrar esta mensaxe

Responder a esta mensaxe
Autor: catuaba
Data:  
Para: LibreMesh.org project mailing list
Asunto: Re: [lime] QoS in Libremesh
hi Cris,

Thank you for the welcome <3

> I'm helping in this community in Italy in countryside:
> https://antennine.noblogs.org/
> and we don't use piranha, sorry.


it is ok, pirania was the first approach we used to control the network
currently it was turned off.

> but If you have some human-info about your community, the shape, the
> organization, please share some links


Here you can have some info about the network
https://open.janastu.org/projects/cowmesh

We are not having the phone booths anymore, currently we have a kiosk
inside the school to allow who does not have a device to connect to
internet and the "gate" is down, APs connection is open without pirania.
We have a server providing some services but seems that people is not using
it so much, understanding their internet usage could help us to
provide services that feets better to their needs (I believe that
talking to them also can help in such way but I have the perception of when we
directly ask the answer use to be different of how they really behave
hehe)

> And if the community wants.. is also possible appear in the homepage of
> Libremesh website, we are in the process to do.


I spoke to them and they agree, what infos we should to provide?

> tc .. I don't know what are you refer..


this stands for traffic control, it is a package that can be used by
libremesh/openwrt.

> But if QoS is Quality of Service,
> in our mesh network we use prometheus (node-exporter) on each antenna, with
> a rasberrypi with prometheus collector and a graphana istance.
> so we have an idea when the antennas are down, or deaf.
> we started to use prometheus when we took a lot of used devices and we need
> to test.. we documented here
> https://antennine.noblogs.org/post/2021/05/31/test-e-monitoring-delle-antenne/
> but documentation about the modern state of our monitoring.. hum.. at the
> moment maybe we have only code :/
> https://git.lattuga.net/antennine


thank you for sharing the documentation. I was reading it and can you
please help me to understand some details?

I saw that you have configured prometheus with 3 groups:
- station(AP), node(mesh) and home(gatway)
So the configuration of each node-export of them are diffents?
Do you have examples?


>
> because we have 4 mesh networks not joined phisically... so to replicate the
> raspberrypi monitoring configuration, we started to use Ansible to install
> and manage configuration of the 3 raspberry pi (in theory we need 5, one for
> valley and one of test)... more a VPS, virtual private server, that is the
> point of connections of the 4 valleys.. and is the VPS where comes the VPN
> (virtual private network), more a computer that is our phisical server and
> give services over the 2 (but a day will be 4 :) valleys
>
> hum.. BUT... at the moment we turn off our server, because electricity
> become too expensive :/ hemm so is a bit a mess.. and I'm happy to explain
> but there is not yet written documentation because we have to change
> strategy to have a cheaper solution, or find the money.


ohhhhh it is nice to use Ansible to give a hand <3 in our case, is a
single network and we was already using grafana to monitor local services
so adding a new category for the network itself was not an issue BUT
currently all the nodes are threated as the same hehe...
Can you share please the template you are using to plot the graphs? I
would like to compare the differences between what we have.

Was really nice to meet you and everybody else during the meeting.

bjs,
catuaba
--
Where the life can take you?