Ultimate Linux Box: A Work in Progress
Being as LinuxWorld was last week and some of you have seen the box already, I thought I would let the rest of you in on the surprise we have in store and some changes to the design of the Ultimate Linux Box we had to make based on that surprise.
The surprise is this year's ULB will be a dual Opteron workstation. Early rumors said the Opteron boards would end up being AGP-less for a while, but that turned out not to be the case. The current platform is a pair of Opteron 242's atop an Arima workstation board, which contains an 8x AGP slot, 32-bit PCI slots, optional onboard Serial-ATA interfaces, USB-2 ports and a Broadcom gigabit Ethernet port. I haven't yet had a chance to run the benchmark suite I mentioned last time against the machine, but initial indications say the machine is amazingly fast, due in part to the Opteron's HyperTransport direct, dual-channel memory bus. The one kernel compile I tried before giving up the machine for use at LinuxWorld completed in a stunning four minutes. This particular motherboard has a unique setup; instead of mandating that memory be installed in pairs, as do so many other boards, this one accepts a single DIMM in Slot 1. It simply uses only one of the two HyperTransport channels. Of course, this means your memory access speed is halved, but if you're trying to diagnose hardware issues, it sure does make the job easier.
We had to make two changes in light of the new CPU/motherboard combination. The first was the case; the original Lian Li case I described in "A Case Study" would not accommodate the bigger Opteron motherboard. So Monarch, our supplier, went back to Lian Li and got them to build us a longer, custom case based on the PC-6070 case. It lacks the top fan port of the PC-65B we originally reviewed, but it comes with the same soundproofing as the PC-6070 we reviewed in the noise article in addition to quiet fans. It also proves to be even easier to work on than its little brothers, thanks to the extra room and all those thumbscrews. Changing major components such as power supplies, disk arrays and motherboards is a breeze. The case also comes with a cute little ring clamp device that sticks to the inside floor of the case and keeps all those Serial-ATA cables from getting in the way.
The second change we made was a switch in video cards. On July 22, ATI finally released support for the Fire GL cards for XFree86 4.3, but only for 32-bit Linux and still only in RPM format--but that won't matter, as you will see in a few minutes. NVIDIA, on the other hand, released XFree86-version-agnostic support for the AMD-64 on December 11, 2002. So the ULB this year will come with an NVIDIA Quadro card instead of the ATI Fire GL X1. The Quadro card comes with a new X functionality called TwinView, in which both display ports share a single framebuffer and thus can display 3-D accelerated video on multiple displays, unlike Xinerama. Obviously, this requires quite a bit of mucking about with the XF86Config file, but the README NVIDIA publishes seems reasonably straightforward.
Interestingly enough, all three of the major graphics hardware vendors--NVIDIA, ATI and Matrox--have gone proprietary with their drivers. No longer is there a way to get premium 3-D hardware without resorting to something for which you don't have all the source code. This is a pretty sad state of affairs, which is mitigated only by NVIDIA's obvious dedication to staying ahead of the game--a welcome change from a few years ago. I really wish the graphics people would take a cue from 3Ware and Intel: GPL the drivers and concentrate on making the hardware the best it can be. Unfortunately, I don't see that happening anytime soon.
Obviously a 64-bit computer needs a 64-bit operating system. We have selected SuSE's Linux Enterprise Server 8 for this. Yes, I realize this isn't Free as in Beer, because you can't go out and download it. It'll cost you $767.00 according to the SuSE web site at press time. Remember, the GPL specifically allows you to sell GPL software; we learned that earlier this year in the DansGuardian review. But it is still Linux, still GPL, still Free as in Speech, and for your semolians you get a year of basic (e-mail) maintenance and free updates.
You're still in sticker shock.
Well, I tried what passes for a downloadable x86_64 distribution. It shall remain safely nameless, to avoid embarrassing the vendor. I couldn't get X to run at all. A lot of stuff was broken. In short, it was totally unsuited to be associated with the Ultimate Linux Box. So we're going with SuSE, and pay the price for being on the leading, bleeding edge. I believe the Right Stuff quote is, "No bucks, no Buck Rogers." I know, we would like to believe that Linux is where you can get good, fast and cheap, all on a single CD. Actually, that's still true. It's still cheap, it's just not Free as in Beer. Given that distributions for alternate platforms are often quite different from those for Lintel boxes, I'll be reviewing SLES 8 separately in an upcoming column.
So that's what we've been doing with the Ultimate Linux Box--a new case, a new video card and the Opteron, running the first and what so far is the only publicly available 64-bit-native OS for that processor: Linux.
Glenn Stone is a Red Hat Certified Engineer, sysadmin, technical writer, cover model and general Linux flunkie. He has been hand-building computers for fun and profit since 1999, and he is a happy denizen of the Pacific Northwest.
|Dynamic DNS—an Object Lesson in Problem Solving||May 21, 2013|
|Using Salt Stack and Vagrant for Drupal Development||May 20, 2013|
|Making Linux and Android Get Along (It's Not as Hard as It Sounds)||May 16, 2013|
|Drupal Is a Framework: Why Everyone Needs to Understand This||May 15, 2013|
|Home, My Backup Data Center||May 13, 2013|
|Non-Linux FOSS: Seashore||May 10, 2013|
- RSS Feeds
- Making Linux and Android Get Along (It's Not as Hard as It Sounds)
- Using Salt Stack and Vagrant for Drupal Development
- New Products
- Validate an E-Mail Address with PHP, the Right Way
- Drupal Is a Framework: Why Everyone Needs to Understand This
- Download the Free Red Hat White Paper "Using an Open Source Framework to Catch the Bad Guy"
- A Topic for Discussion - Open Source Feature-Richness?
- Dynamic DNS—an Object Lesson in Problem Solving
- Home, My Backup Data Center
- Please correct the URL for Salt Stack's web site
1 hour 59 min ago
- Android is Linux -- why no better inter-operation
4 hours 14 min ago
- Connecting Android device to desktop Linux via USB
4 hours 43 min ago
- Find new cell phone and tablet pc
5 hours 41 min ago
7 hours 10 min ago
- Automatically updating Guest Additions
8 hours 18 min ago
- I like your topic on android
9 hours 5 min ago
- This is the easiest tutorial
15 hours 40 min ago
- Ahh, the Koolaid.
21 hours 19 min ago
- git-annex assistant
1 day 3 hours ago
Enter to Win an Adafruit Pi Cobbler Breakout Kit for Raspberry Pi
It's Raspberry Pi month at Linux Journal. Each week in May, Adafruit will be giving away a Pi-related prize to a lucky, randomly drawn LJ reader. Winners will be announced weekly.
Fill out the fields below to enter to win this week's prize-- a Pi Cobbler Breakout Kit for Raspberry Pi.
Congratulations to our winners so far:
- 5-8-13, Pi Starter Pack: Jack Davis
- 5-15-13, Pi Model B 512MB RAM: Patrick Dunn
- 5-21-13, Prototyping Pi Plate Kit: Philip Kirby
- Next winner announced on 5-27-13!
Free Webinar: Hadoop
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?