<div dir="ltr"><div>I stand by what Joe says. Good advice.</div><div><br></div><div>When at Viglen XMA we had a similar system for field testing and burnin.</div><div>However we did not use a USB bootable system.</div><div>We would take a complete head node on site when I first joined the company, using Rocks Linux.</div><div>I am no far of Rocks I'm afraid - sorry Rocks guys but it just got under my skin.</div><div><br></div><div>In later days we got some small form factor HP systems and I used an OpenHPC install on them.</div><div>If you get a small form factor system with a PCI Express slot this can use a 10Gbps or an Infiniband card so you have the correct network to PXE over.</div><div><br></div><div>So I would look at a portable small form factor system, using OpenHPC or the utilities with Joe suggests.</div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr">On Fri, 11 Jan 2019 at 16:07, Joe Landman <<a href="mailto:joe.landman@gmail.com">joe.landman@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;padding-left:1ex;border-left-color:rgb(204,204,204);border-left-width:1px;border-left-style:solid"><br>
On 1/11/19 7:59 AM, Richard Chang wrote:<br>
> Hi,<br>
> I would like to know if we have or can make( or prepare) a USB <br>
> bootable OS that we can boot in a cluster and its nodes to test all <br>
> its functionality.<br>
><br>
> The purpose of this is to boot a new or existing cluster to check its <br>
> health, including Infiniband network, any cards, local hard disks, <br>
> memory etc, so that I don't have to disturb the existing OS and its <br>
> configuration.<br>
><br>
> If possible, it would be nice to boot the compute nodes from the <br>
> master node.<br>
><br>
> Anyone knows of any pre-existing distribution that will do the job ? <br>
> Or know how to do it with Centos or Ubuntu ?<br>
<br>
FWIW: this is one of the uses cases of <br>
<a href="https://github.com/joelandman/nyble" target="_blank" rel="noreferrer">https://github.com/joelandman/nyble</a> . It works with CentOS, Debian, and <br>
Ubuntu (though I've not pushed the 18.04.1 changes yet).<br>
<br>
I have a rudimentary USB target I was going to clean up soon, and the <br>
images can be centrally booted from a pxe server, and pull/run scripts <br>
post boot.<br>
<br>
Runs in RAM, you can modify the distributions to your hearts content. I <br>
have a few private repos here which have NVidia + MLNX + other drivers <br>
and related bits already built in.<br>
<br>
I've set up many systems with this, tying it together with <br>
<a href="https://github.com/joelandman/tiburon" target="_blank" rel="noreferrer">https://github.com/joelandman/tiburon</a> for boot control. This was <br>
originally used at Scalable Informatics when we were alive, and has <br>
evolved significantly since then.<br>
<br>
If you want a simple pure USB distro for this, try SystemRescueCD, <br>
though I don't think it does Infiniband, or most drivers.<br>
<br>
<br>
-- <br>
<br>
Joe Landman<br>
e: <a href="mailto:joe.landman@gmail.com" target="_blank">joe.landman@gmail.com</a><br>
t: @hpcjoe<br>
w: <a href="https://scalability.org" target="_blank" rel="noreferrer">https://scalability.org</a><br>
g: <a href="https://github.com/joelandman" target="_blank" rel="noreferrer">https://github.com/joelandman</a><br>
l: <a href="https://www.linkedin.com/in/joelandman" target="_blank" rel="noreferrer">https://www.linkedin.com/in/joelandman</a><br>
<br>
_______________________________________________<br>
Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org" target="_blank">Beowulf@beowulf.org</a> sponsored by Penguin Computing<br>
To change your subscription (digest mode or unsubscribe) visit <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank" rel="noreferrer">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
</blockquote></div>