Setting Up a SPARCstation
Many PC users have taken the plunge into the Unix world with Linux and, I imagine, quite a few Unix die-hards like me have taken their first, faltering steps with PC hardware, using Linux, too. I have to say that for a long-time Sun user, the attempted transition to PCs was not without its trials. The mysterious workings of BIOS and extended-vs-expanded memory still elude me, so it was with more than a little relief that I noted, towards the end of last year, the release of Red Hat 4.0 for the SPARC architecture. Now we Sunnites can run our favourite OS on our favourite hardware, and the PC folks have a chance to run a familiar OS on what many may have, until recently, considered rather exotic hardware. Many institutions are selling off, or even scrapping, their first generation SPARCs as they become less and less useful when running the relatively “heavy” OS which Sun now ships. Many of these workhorses are becoming available to the individual user at little or no cost, and Linux provides an ideal solution to the question of which OS to use.
SPARCstations generally offer better performance than a PC of equivalent age and come with a relatively standard set of peripherals and interfaces. Among the benefits are a much higher screen resolution than the old VGA-based 386 machines (the standard resolution for Suns is 1152x900) and built-in audio, Ethernet and SCSI controllers.
In this article I hope to introduce those readers unfamiliar with Sun hardware to a few of the peculiarities of the breed, provide pointers to which systems are good buys and which should be avoided, and to provide enough information for you to get Red Hat's SPARC distribution up and running on your machine.
What systems are out there and what should you look for when shopping for a Sun workstation on which to run Linux?
First of all, several groups of machines are not currently supported under Red Hat Linux/SPARC and will not work under the 4.0 or 4.1 releases.
None of the VME-based deskside or server machines are supported (i.e., the 400 and 600 series).
The SS1000 and SS2000 servers are not supported.
The newer “Ultra” based machines are not supported.
I should emphasize here that SPARC Linux is one of the platforms where changes are happening very rapidly (kudos to the dedicated band of portmeisters), and I expect that by the time that you actually read this an Ultra version will be available. However, the Red Hat version 4.0 and 4.1 releases discussed here do not support that architecture.
More recent desktop machines, the Classic and SPARCstations 10 and 20 (all “4m” architecture), run SPARC Linux quite happily, and the testing and boot details given later in this article are generally valid for them. However, it's unlikely that the average home user will come across one of these machines on the market at a price which he'd be willing to pay for an unfamiliar piece of hardware, so I'd like to concentrate here on the machines which you're most likely to encounter, the older, “4c” architecture systems.
First, you need to be aware that several of Sun's chassis will accommodate several different types of CPU. A good example of this is the original “pizza box” SPARCstation enclosure (about the same size as a large size pizza delivery box, with a distinctive “dimple” pattern on the front and cooling holes in the same pattern on the sides). This chassis was virtually unchanged between several different models, the SPARCstation 1, 1+ and 2 (the SPARCstation 2 had a small fan and a floor grille added for disk cooling).
Unfortunately for potential buyers, this chassis also accommodates the Sun 3/80 CPU. The 3/80 is 68030 based (not SPARC) and, at the moment, no Linux port is available for it. Unless you're intent on joining the band of volunteers working on the Sun3 porting project, you do not want to buy a 3/80 in SPARCstation clothing. Don't accept that the logo on the front of the machine actually reflects what is inside. Take a good close look at the business end of the CPU. Even if you can't open the case to check the CPU chip, a dead give-away that the machine is actually a 3/80 is the presence of a 9-pin D-type serial port on the CPU. None of the desktop SPARCstations has this type of connector, although most, including the 3/80, have a 15-pin D-type Ethernet connector, so count those pins.
Any of the other machines mentioned above, the SPARCstation 1, 1+ or 2, will run Linux quite happily. They generally come configured with a floppy drive and one or two internal hard disk drives. They don't have an on-board frame-buffer though and one of the three available SBus card slots must be sacrificed to add one. Note: on the 1 and 1+ machines, the third SBus slot is marked as a “slave” slot and is only suitable for frame-buffer boards. It will not support I/O cards such as SCSI or Ethernet.
Two other models appearing on the second-user market more frequently nowadays are the SPARCstation SLC and ELC systems. Both of these machines are easily mistaken as being nothing more than ordinary monochrome monitors, as the CPU is built into the housing of a 17 inch grey scale monitor. The clues are the SCSI, RS232 and keyboard connectors on the back panel of the machine. Keep an eye open for these two, and you could pick up a bargain.
Although both models first saw the light of day after the original SPARCstation 1, they were explicitly targeted at the bottom end of the market and mostly ended up as diskless workstations. The ELC is the more powerful of the pair, though not by much, and it can be recognized by the CPU access panel at the top rear of the monitor housing. However, as with the previously mentioned machines, CPUs and chassis are interchangeable, so the only way to tell for sure which machine you have is to power it on.
Another two systems with a common chassis are the IPC and the IPX. The enclosure for these two is smaller in width and depth and a little taller than the “pizza box” machines, and it is designed to have the same desktop “footprint” as the (separate) monitor. This enclosure can be installed under the monitor or stood on edge using an “L” shaped, blue-grey plastic stand supplied by Sun along with the system.
The compact design of the enclosure emphasizes the fact that these systems were targeted at the desktop user. The chassis is limited to a single 3.5 inch hard drive bay (plus a floppy), and non-standard mini-DIN style sockets are used for the RS232 connectors on the CPU back-panel.
Both machines have a built-in frame-buffer. The on-board frame-buffer supplied with the IPC is only monochrome, although it is possible to add colour by utilizing one of the two available SBus slots. The IPX sports a colour frame-buffer and uses a different type of SIMM, making it possible for the IPX to have a greater overall memory capacity while actually having only one third of the number of SIMM sockets of the IPC. Of the two, the IPX is also the more powerful.
|Non-Linux FOSS: libnotify, OS X Style||Jun 18, 2013|
|Containers—Not Virtual Machines—Are the Future Cloud||Jun 17, 2013|
|Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer||Jun 12, 2013|
|Weechat, Irssi's Little Brother||Jun 11, 2013|
|One Tail Just Isn't Enough||Jun 07, 2013|
|Introduction to MapReduce with Hadoop on Linux||Jun 05, 2013|
- Containers—Not Virtual Machines—Are the Future Cloud
- Non-Linux FOSS: libnotify, OS X Style
- Linux Systems Administrator
- Validate an E-Mail Address with PHP, the Right Way
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Senior Perl Developer
- Technical Support Rep
- UX Designer
- Introduction to MapReduce with Hadoop on Linux
- RSS Feeds
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?