[HH] Has anyone seen a solution to this?

[Kurt Keville]

Well the first batch of nodes we are going to try this with are the Jetson
TK1s...
http://www.nvidia.com/object/jetson-tk1-embedded-dev-kit.html    ... there
aren't a lot of I/O options and we have already tested the onboard GigE and
found it to be lacking in a couple of areas. We were thinking it would be
good for intracluster commo but not a lot of sustained writes to the head
node. NVidia understandably didn't put a lot of money into the NIC on board
the Jetson so it isn't much to write about. There are some tantalizing GPIO
pins that might represent a path to a better network although a lit search
doesn't turn much up. The next gen Jetson model (the TX1) has a GPIO pinout
that is compatible with Ras Pi Shield standard but I don't think that is
true with the Tegra K1...  That's how I ended up investigating this path;
bandwidth to local disk is pretty good on these dev kits and there aren't
many (better, faster, cheaper) ways to get data off these... the best
picture of the µcluster so far is at ...
http://web.mit.edu/~kkeville/www/Jetson3.jpg

On Mon, Feb 15, 2016 at 12:00 PM, Mark Komarinski <mkomarinski at wayga.org>
wrote:

>
> On Feb 14, 2016, at 4:21 PM, Kurt Keville <kurt.keville at gmail.com> wrote:
>>
>> I am looking at putting a cluster together, say 6 nodes with local disk
>> and 1 GigE network port... but instead of having them talk to the head node
>> via network I want the head node to have 6 connections directly to the hard
>> drives... I figure this would require some sort of SATA pigtail in the
>> client nodes, and a 6 node SATA card in the head node, but is there a way
>> to run an arbitrator on the head node that won't interfere with the
>> operation of the local disk, since the nodes wouldn't necessarily know
>> about the head node... this way I remove the need to copy files to the head
>> node via GigE...
>>
>> I know I can do parts of this independently... I have a couple of JBOD
>> boxes here and as long as I can just let the head node know it is a
>> read-only filesystem I might be in business... especially if I can ensure
>> that I am not doing head node reads at the same time as client node
>> writes...
>>
>>
> Hi Kurt,
>
> Rather than look at the solution you're proposing, I’d like to look at the
> problem and understand it.  What are you trying to solve?  It sounds like
> you have data on the head node that you’d like to get to the compute nodes
> but not go over the GigE port (why avid GigE?).  Do the compute nodes need
> to send data back to the head node?  Is it the same data on all 7 hard
> drives?
>
> I’ve got a bunch of ideas on how you can solve this, but I don’t think
> that having multiple SATA cables per drive is the best way to go.
>
> -Mark
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.blu.org/pipermail/hardwarehacking/attachments/20160215/48b9cade/attachment.html>