Skip to Content.
Sympa Menu

perfsonar-user - RE: [perfsonar-user] Inexpensive testing nodes - LIVA in particular hardware questions

Subject: perfSONAR User Q&A and Other Discussion

List archive

RE: [perfsonar-user] Inexpensive testing nodes - LIVA in particular hardware questions


Chronological Thread 
  • From: "Chevalier, Scott S" <>
  • To: Daniel Doyle <>, Hyojoon Kim <>
  • Cc: Larry Blunk <>, "" <>
  • Subject: RE: [perfsonar-user] Inexpensive testing nodes - LIVA in particular hardware questions
  • Date: Tue, 8 Dec 2015 23:12:19 +0000
  • Accept-language: en-US

Joon,

 

As Dan has already mentioned the 2G Liva-X model should be able to perform as a 1G testpoint. My personal experience with the LIVA’s have been with the base BAT-mini model and our internal testing shows those testing at around 940Mbps.

 

There are several community members using the BAT-mini LIVAs with varying degrees of success. Personally, I have had a failure rate of 2 per 15 or so and have had some other production issues; power button/boot-up instability for instance. The LIVA boxes are also limited to the Debian/Ubuntu pS packages. (http://docs.perfsonar.net/low_cost_nodes.html)

 

With that in mind there are other devices which have been under-going some limited testing and showing great promise, in the same price range as the 4G LIVA-X model. The second generation NUC and the Zotac Zbox are really solid platforms, offering additional port availability, expandable memory and drive options (the base node does not come with memory/drive, so additional shopping is required), and OS flexibility (able to use the more community-supported CentOS ISO packages).

 

Regardless of which devices you end up choosing, please let us know how the small nodes end up working for your deployment and any issues you encounter along the way.

 

Thanks,

 

Scott Chevalier

Network Systems Analyst

International Networking @ IU

812-856-9964

 

 

 

From: [mailto:] On Behalf Of Daniel Doyle
Sent: Tuesday, December 08, 2015 5:13 PM
To: Hyojoon Kim <>
Cc: Larry Blunk <>;
Subject: Re: [perfsonar-user] Inexpensive testing nodes - LIVA in particular hardware questions

 

Joon,

 

From my own experiences, I have been using the 2G LIVA X model for testing and it seems to work just fine. The 4G model might be more futureproof (and I think it comes with more disk space, which might be relevant if you're doing more), but these are unlikely to be long long term things anyway and will be superseded at some point in the future by some new shiny. Since these nodes aren't intended to be running the MA or the UI or anything their resource needs are rather modest by today's standards.

 

Others' opinions may vary, but I'd go with the cheaper of the two especially if you're getting multiples.

 

My 2c,

-Dan

 

On Dec 8, 2015, at 4:39 PM, Hyojoon Kim <> wrote:

 

Hello,

Related to this, is anyone aware of any big difference in terms of perfSONAR measuring capability between LIVA X 2GB vs. LIVA X 4GB Mini PC?

The 2GB memory LIVA X is $104 while the 4GB memory LIVA X is $144 on newegg, so if there is no significant advantage of using 4GB LIVA X instead of the 2GB LIVA X, we want to stick with the 2GB one (and I think the community mostly tested with a 2GB version.)

Any experience or thoughts to share?

Thanks,
Joon


On Oct 16, 2015, at 3:18 PM, Larry Blunk <
<>> wrote:

Signed PGP part

Looks like the LIVA X is currently only $10 more on Newegg--

http://www.newegg.com/Product/Product.aspx?Item=N82E16856501011

Specs show a slightly faster CPU vs. LIVA (N2808 vs. N2807) plus
an extra USB port.  It's also pre-assembled vs. a kit which is a plus.

One thing to note is that Ubuntu enables the "ondemand" init script
by default which puts the CPU in "powersave" mode.   In testing the
LIVA, I found that this seems to limit throughput a bit in performance
tests.  I get around 900Mbps TCP throughput vs. 940Mbps in
"performance" mode.  Also saw some packet loss when doing UDP tests
with 1500 byte datagrams at 1 Gbps in powersave mode.  You can disable
the script with the following which will leave the CPU in performance
mode.

update-rc.d -f ondemand remove



-Larry Blunk
Merit


On 10/13/2015 01:21 PM, Daniel Doyle wrote:
Hi Casey,

The Liva Black (or BAT?) was the original tested hardware platform.
This one in particular (not trying to promote newegg as a vendor,
just handy link):
http://www.newegg.com/Product/Product.aspx?Item=N82E16856501007.
For the most part it just worked with Ubuntu 12.05 Desktop,
although a couple of models had a few hardware quirks.

At IU we're going to be looking at the Liva X as a new release of
that hardware. There is actually one sitting on my desk right now.
You can get these from amazon or newegg or whatever today, although
to the best of my knowledge nobody has done any testing on them
just yet. If you want one to tinker go ahead, but I might hold off
on ordering 1000.

-Dan

On Oct 13, 2015, at 1:12 PM, Casey Russell <>
<>> wrote:

Group,

We're interested in obtaining a new small node hardware platform
for testing.  I was interested particularly in the LIVA mini-pc.
They were mentioned at TechEx and are mentioned on the small node
page off the main PS webpage.

What I'm interested in is which of the LIVA mini-PCs has been
tested by I2 or ES.net<http://es.net/> <http://ES.net<http://es.net/>> (or others) with
PerfSonar? The Liva?  Liva X?  Liva X2?  Liva Core?  I believe
what I saw talked about at TechEx was either the Liva or the Liva
X, probably the Liva X.

I also remember talk of a new hardware revision about to be
released for these, does anyone know when that should drop
(roughly), so I'll know whether to buy this week or wait a couple
of weeks?

Thank you

Casey Russell Network Engineer Kansas Research and Education
Network 2029 Becker Drive, Suite 282 Lawrence, KS  66047
(785)856-9820  ext 9809 <>
<>

Dan Doyle GlobalNOC Software Developer 1-812-856-3892
<> <>

 

Dan Doyle

GlobalNOC Software Developer

1-812-856-3892

 

 

 




Archive powered by MHonArc 2.6.16.

Top of Page