[Bismark-devel] assembling a testbed

[Bismark-devel] assembling a testbed

[Dave Taht]

[-- Attachment #1: Type: text/plain, Size: 925 bytes --]

I had rather hoped the wndr3700v2 routers would have arrived by now. Can you
ping them to see if they will arrive in time? If not, acquiring sufficient
routers for the trip is becoming rather important. I'd like to be testing by
wednesday.

Given that we had over 10% infant mortality on the 2.4Ghz radios, I worry.
Also since the problem has been the same across multiple wndr3700s I figure
it's a manufacturer switch/mac interface change that can be fixed (somehow),
with a bit of work at the driver level (might require dissassembly, too)

Lacking a testbed as such, I have been assembling the 3 nano-5 routers I
have into a form that could be used to exercise the few wndr3700v2s we have.
It would be good to coax as many people in the lab as possible to login via
the testbed routers during the test period, as well.

-- 
Dave Täht
SKYPE: davetaht
US Tel: 1-239-829-5608
http://the-edge.blogspot.com

[-- Attachment #2: Type: text/html, Size: 1028 bytes --]

Re: [Bismark-devel] assembling a testbed

[Nick Feamster]

We have been working on getting them ASAP.  Shipment was delayed for bureaucratic reasons. :-(  But, we were in the purchasing office this morning, and hopefully we'll have them soon.  Fallback plan is a shopping spree.

-Nick

On May 9, 2011, at 3:04 PM, Dave Taht wrote:

> I had rather hoped the wndr3700v2 routers would have arrived by now. Can you ping them to see if they will arrive in time? If not, acquiring sufficient routers for the trip is becoming rather important. I'd like to be testing by wednesday. 
> 
> Given that we had over 10% infant mortality on the 2.4Ghz radios, I worry. Also since the problem has been the same across multiple wndr3700s I figure it's a manufacturer switch/mac interface change that can be fixed (somehow), with a bit of work at the driver level (might require dissassembly, too)
> 
> Lacking a testbed as such, I have been assembling the 3 nano-5 routers I have into a form that could be used to exercise the few wndr3700v2s we have. It would be good to coax as many people in the lab as possible to login via the testbed routers during the test period, as well. 
> 
> -- 
> Dave Täht
> SKYPE: davetaht
> US Tel: 1-239-829-5608
> http://the-edge.blogspot.com 
> _______________________________________________
> Bismark-devel mailing list
> Bismark-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bismark-devel

Re: [Bismark-devel] assembling a testbed

[Dave Taht]

[-- Attachment #1: Type: text/plain, Size: 2026 bytes --]

On Mon, May 9, 2011 at 1:06 PM, Nick Feamster <feamster@cc.gatech.edu>wrote:

> We have been working on getting them ASAP.  Shipment was delayed for
> bureaucratic reasons. :-(  But, we were in the purchasing office this
> morning, and hopefully we'll have them soon.  Fallback plan is a shopping
> spree.
>

My biggest fear is that since we are getting all of them in a batch instead
of mixed up on a shelf, and my diagnosis of the infant mortality is an
unknown mac/switch change...

 that 100% OR *0*% of the soon to arrive routers will work out of the box.

This sort of thing costs me sleep. I wish I had one of the borked routers to
at least poke at the phy a bit...


>
> -Nick
>
> On May 9, 2011, at 3:04 PM, Dave Taht wrote:
>
> > I had rather hoped the wndr3700v2 routers would have arrived by now. Can
> you ping them to see if they will arrive in time? If not, acquiring
> sufficient routers for the trip is becoming rather important. I'd like to be
> testing by wednesday.
> >
> > Given that we had over 10% infant mortality on the 2.4Ghz radios, I
> worry. Also since the problem has been the same across multiple wndr3700s I
> figure it's a manufacturer switch/mac interface change that can be fixed
> (somehow), with a bit of work at the driver level (might require
> dissassembly, too)
> >
> > Lacking a testbed as such, I have been assembling the 3 nano-5 routers I
> have into a form that could be used to exercise the few wndr3700v2s we have.
> It would be good to coax as many people in the lab as possible to login via
> the testbed routers during the test period, as well.
> >
> > --
> > Dave Täht
> > SKYPE: davetaht
> > US Tel: 1-239-829-5608
> > http://the-edge.blogspot.com
> > _______________________________________________
> > Bismark-devel mailing list
> > Bismark-devel@lists.bufferbloat.net
> > https://lists.bufferbloat.net/listinfo/bismark-devel
>
>


-- 
Dave Täht
SKYPE: davetaht
US Tel: 1-239-829-5608
http://the-edge.blogspot.com

[-- Attachment #2: Type: text/html, Size: 2907 bytes --]

Re: [Bismark-devel] assembling a testbed

[Nick Feamster]

On May 9, 2011, at 3:14 PM, Dave Taht wrote:

> My biggest fear is that since we are getting all of them in a batch instead of mixed up on a shelf, and my diagnosis of the infant mortality is an unknown mac/switch change...
> 
>  that 100% OR *0*% of the soon to arrive routers will work out of the box.
> 
> This sort of thing costs me sleep. I wish I had one of the borked routers to at least poke at the phy a bit...

We will do a shopping spree if need be.

Re: [Bismark-devel] assembling a testbed

[Dave Taht]

[-- Attachment #1: Type: text/plain, Size: 501 bytes --]

From a testbed perspective, it would be good to drive a LOT of traffic
through the routers of all sorts, while watching bismark go.

So one thought is to drive them from an existing box(es), close by, from
multiple network cards from a given machine or set of machines.

Is there a machine available that has a few spare PCI or ePCI slots?

are there some extra network cards lying around? gigE pci and pci-e cards
are very cheap nowadays and there are also wireless pci adaptors of various
sorts....

[-- Attachment #2: Type: text/html, Size: 530 bytes --]