[Rack] Networking as of 2013-10-02

Omar Zouai ozouai.bavc at gmail.com
Tue Oct 8 16:31:12 UTC 2013


I believe I did look at the right box, correct me if I'm wrong. A green box
with 4 LAN ports. 3 LEDs, green for power, yellow for disk, red for error.
Console port, green antennae. All the LAN ports went back into the juniper
switch.

Swap out a small router for a 150w server? Yep, that's my idea in a
nutshell. Even though it might be more resource intensive, there's always
pros and cons to an idea.  Here's some that I see.
Pros:
    Able to incorporate gateway, dhcp, dns into a singe box.
    Able to shape bandwidth to prevent hogging.
    Able to create different networks for different clients.
    Able to efficiently blacklist anything that would make the FBI storm
into NoiseBridge.
    Logging of how the network is being used.
    High memory/cpu.
Cons:
    Uses a lot more power
    Space consuming.
    More sensitive.
    Less redundancy.

I only see a few key cons, but that's limited to my perspective.
On Oct 7, 2013 12:19 PM, "Ben Kochie" <ben at nerp.net> wrote:

> On Mon, 7 Oct 2013, Omar Zouai wrote:
>
>
>> I did consult the rack. Hence this current debate. Yes, when the network
>> went down, the r00ter
>> was rebooted multiple times. And yes, I let it sit for half an hour, both
>> on and off. I did
>> what was necessary to maintain a internet connection at the time in the
>> space, and came up with
>> an idea on how it can be improved.
>>
>> Yes, I did read the damned wiki while trying to get the internet back
>> up(mobile data).
>>
>> CF Card? After 4 unsuccessful boots, I opened r00ter up. There wasn't a
>> CompactFlash in it.,
>> the only thing remotely visible to it is the WiFi chipset. Maybe this is
>> a problem?
>>
>
> Are you sure you looked at the right box?
>
>  Correction, Networking is simple, but will get more and more complex;
>> which would be fine if
>> all the equipment was "perfect", and did everything it was supposed to do
>> without errors and
>> never failed.  But we don't live in a perfect world, things will have
>> errors, and equipment
>> will go offline. That's where the nightmare comes in.
>>
>
> No, it was already as complex as it needed to be.  We had a nice simple
> all-solid-state router to connect, failover, and load-balance our two ISPs.
>  You want to replace this with a 150W server with fans and hard drives?
>  And you call this "simpler"?
>
> We even simplified the network just a few months ago by removing two
> switches from the critical path and replaced it with one good Juniper
> switch.
>
> The noisebridge network is actively maintained by this group.
>
>  I highly doubt watching the rack is good enough to prevent fuckery. If
>> someone was determined
>> enough to do something stupid, they simply would just go do it, without
>> consulting the Wiki or
>> Rack.
>>
>> On Oct 7, 2013 3:53 AM, "Ben Kochie" <ben at nerp.net> wrote:
>>
>>
>>                   The network at noisebridge is not a toy for you to play
>>                   with.
>>
>>
>>             Disclaimer: I actually agree that it would be much more
>> convenient if
>>             the network wasn't dramatically gutted when not necessary. I
>> only point
>>             out the following technicality.
>>
>>
>>             This is "doacracy" at its best (and worst). Technically, it
>> is a toy
>>             that people could play with. It's a shame the play wasn't
>> "playing
>>             nicely" with the work already done (and previously working).
>>
>>
>>       Actually, the network is one of the few places that is not a toy to
>> play with.  We
>>       (the rack list) keep an eye on things and try and keep the fuckery
>> on the network
>>       to a minimum.
>>
>>             Glen
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.noisebridge.net/pipermail/rack/attachments/20131008/84d3b257/attachment.html>


More information about the Rack mailing list