Tuesday, November 18, 2008

Sun Unified Storage 7000

I was at an internal training all week last week and didn't get a chance to talk about our new Storage products.  I know, I know, Stop laughing.  And yes I did say Storage products.  And no, they are not OEM'ed and YES they do work.

Don't believe me.  Try it out yourself.  Go here and download the VMWare image and run it on your laptop or vmware farm.

Let me know what you think.  I started the VMWare image and it asked me for a password, so I gave it, and it booted up.  I entered the website it gave me, and logged in with root and then the password I gave it.  No issues.  I then added the 15 x2GB drives that it came setup with and created a pool.  After that I added a mount point and shared out an NFS point that I was writing to it on my MAC SWEET.

Then I, started writing data to it to see how that dashboard and anyalitics worked.  It was awesome, as I could see which files were being written to and how long it took just at a glance. 

Don't beleive me, try it for yourself.  It's only a 500MB download.

Did you miss the annoucement?

One great review is here

A quote from the peice, which is my favorite so far:

At the same time, the new capabilities offered with products like
the Sun 7000-series storage appliances are paving the way for new takes
on old ideas, such as video-on-demand. Let's just say that a 48TB 7210
can store quite a few HD movies, and there's no such thing as an
overabundance of scratch space when working with audio and video
projects.

So, in the immortal words of Kent Brockman, I, for one, welcome our new storage overlords.







T

Monday, November 17, 2008

Sun layoffs

Well, since everyone is asking, I thought I would post on this.  The full article is here
Bottom line, Sun is right sizing.  We have made a ton of puchases and
need to right size.  I think the below from the article Sum's it up.





Andrew Kotarba, president and CEO of Dewpoint, a Lansing, Mich.-based
solution provider and Sun partner, said that while he hates to see
6,000 people lose their jobs, the Sun partner community agrees it needs
to be done.


"You have a company that still has a healthy revenue run rate,"
Kotarba said. "It needs to get profitable. It gets the pressure off
their back from the earnings report."


Kotarba, too, sees a big opportunity for the channel in the
layoffs. "Sun placed bets on its channel partners to go to market for
them," he said. "With the cuts, they will depend on us more. Dewpoint
is heavily invested in Sun. This is an opportunity for us to do more."


Whether the cuts are enough remains to be seen, however. "Sun
said it will save $700 million to $800 million annually," Kotarba said.
"If there are no other serious downturns, it should be enough. We all
go through this as business owners. We need revenue to be higher than
expenses."


Kotarba said a lot of the questions about Sun's viability are
actually started by competitors, but that Sun's balance sheet and
technology show it to be a very viable company.


"Sun's technology is second to none," he said.
"Some customers ask about it, some don't. We don't have that discussion
with Sun's installed base of customers. They know Sun and its
technology. But when trying to displaceDell (NSDQ: Dell) or HP (NYSE:HPQ) or IBM (NYSE:IBM),
you need firepower. With all the noise in the marketplace, it's hard.
We need the merits of the technology to stand on their own." 





Is SUN going any where, I don't think so.  Though we have to make a profit.  That's what we are trying to do. 





So
that's my thoughts as well.  We have a great Free software stack.  I've
been called by many customers, and in the first breath they ask if I'm
ok, and in the second they ask if we can help them save money cause
they are doing Layoff's as well. :-}


Yes we can:



  1. Glass Fish Application Server $7,875 for Platinum Plus support per 4 sockets, not cores.

  2. Mysql 5K for Mysql Enterprise per Server

  3. Amber Road - Unified Storage - 1-7 Dollars per GB or less


Can we help you save money?  YES WE CAN, just ask us.


 T

Wednesday, November 12, 2008

Linux Customer Adovating Solaris and zfs

http://blogs.smugmug.com/don/2008/10/10/success-with-opensolaris-zfs-mysql-in-production/


 This is a customer that took OpenSolaris and tested compression and zfs and mysql in production.


 Great read for you Linux Lovers.


 T


Friday, November 7, 2008

More Sun Ray Bandwidth

I've been doing some testing this morning with Bandwidth over Sun Ray's.  YOUR RESULTS WILL VARY, THESE ARE JUST AN EXAMPLE, USING MY APPLICATIONS.


First, I forced Sun Ray to 1024x768@60hz


/opt/SUNWut/sbin/utresadm -a -c -t default 1024x768@60


the -t default means no matter what token is put in, default the resolution to 1024x768@60


using SRSS 4.1.


Test 1:  forced UTTSC with -A 8 to force 8 bit windows mode

 lost      0/00% pkts     93 cpu   0% kbytes     35 0.014 Mbps 1.5(2.0) ms
 lost      0/00% pkts     97 cpu   0% kbytes     36 0.014 Mbps 2.0(2.0) ms
 lost      0/00% pkts     92 cpu   0% kbytes     35 0.014 Mbps 1.4(1.7) ms
 lost      0/00% pkts     94 cpu   0% kbytes     35 0.014 Mbps 1.7(1.7) ms
 lost      0/00% pkts     95 cpu   0% kbytes     35 0.014 Mbps 1.5(1.6) ms
 lost      0/00% pkts     93 cpu   0% kbytes     35 0.014 Mbps 1.5(1.5) ms
 lost      0/00% pkts     93 cpu   3% kbytes     35 0.014 Mbps 2.1(1.8) ms


Test 2: Forced UTTSC with a -A 15 No Compression forced on the DTU


lost      0/00% pkts    873 cpu   0% kbytes    779 0.060 Mbps 5.1(5.1) ms
 lost      0/00% pkts    158 cpu   4% kbytes    144 0.056 Mbps 2.0(3.5) ms
 lost      0/00% pkts    157 cpu   0% kbytes    144 0.056 Mbps 2.0(2.8) ms
 lost      0/00% pkts    130 cpu   0% kbytes    135 0.053 Mbps 1.5(2.2) ms


Test 3: Forced UTTSC with a -A 15 and Forced compression on the DTU


 lost      0/00% pkts     60 cpu   0% kbytes     34 0.013 Mbps 4.5(4.5) ms
 lost      0/00% pkts     60 cpu   0% kbytes     34 0.013 Mbps 1.9(3.2) ms
 lost      0/00% pkts    106 cpu   0% kbytes     39 0.016 Mbps 2.2(2.7) ms
-------
 lost      0/00% pkts    458 cpu   0% kbytes    203 0.016 Mbps 1.6(1.6) ms
 lost      0/00% pkts    104 cpu   0% kbytes     38 0.015 Mbps 1.6(1.6) ms


Test 4: Forced UTTSC with a -A 16 and Forced Compression on the DTU


 lost      0/00% pkts    107 cpu   0% kbytes     38 0.015 Mbps 2.0(2.0) ms
 lost      0/00% pkts    104 cpu   0% kbytes     38 0.015 Mbps 1.5(1.7) ms
 lost      0/00% pkts    106 cpu   4% kbytes     38 0.015 Mbps 1.5(1.6) ms
-------
 lost      0/00% pkts    592 cpu   1% kbytes    254 0.019 Mbps 1.5(1.5) ms
 lost      0/00% pkts    106 cpu   0% kbytes     38 0.015 Mbps 1.5(1.5) ms


Test 4: Forced UTTSC with a -A 16 and No compressions on the DTU


lost      0/00% pkts    100 cpu   0% kbytes     98 0.038 Mbps 3.4(2.7) ms
 lost      0/00% pkts    101 cpu   4% kbytes     98 0.038 Mbps 2.9(2.8) ms
 lost      0/00% pkts    100 cpu   0% kbytes     98 0.038 Mbps 1.9(2.3) ms
 lost      0/00% pkts    110 cpu   0% kbytes     94 0.037 Mbps 2.1(2.0) ms


I stopped here, as I would never do 24 bit unless you twisted my arm, I feel there is very little need for an average user to use 24bit color depth.


 What does this all mean?  The protocol that's being used with the Sun Ray's is an adaptive protocol.  I like to describe it with a toll booth.  Let's say I have 8 lanes to pay the toll's and no one else on the toll way.  I can use all 8 lanes if necessary, but as more traffic is added to the toll way, the ability to use all 8 lanes decrease, so the protocol adapts itself to it's current setting. 


The problem with this is 2 fold: 1) if you are testing on a private LAN, you will typically always have higher bandwidth numbers than if your LAN was being fully utilized.  In other words, when test a couple of Sun Ray's and you see 38Kbps, you can't take that number and mutiple it by the number of Sun Ray's being used at once, because our protocol will adapt as the network get's more utilized. 2) if you have a SRSS server on the other side of a WAN from your Sun Ray, the Protocol has a tendency not to approximate the bandwidth correctly, because it may see a very quick LAN and not the slow WAN.


 Solution: in the 4.1SRSS DTU firmware, there is a force compression option in the advanced menu.  You can also force from the params file, but I don't know how to do that yet, and is next on the list.