Operating the ULB: SLES 8 on the Ultimate Linux Box
So the Ultimate Linux Box is back in-house after a brief hiatus to be shown off at LinuxWorld. In the last article, I told you that because of a really bad experience with another, safely nameless distribution on the 64-bit platform, we would be going with SuSE Linux Enterprise Server 8, a pay-for-play package. Because it makes up a significant percentage of the cost of the ULB, I thought it deserved its own review.
Instead of the usual box of books and CD cases thing I've seen from every other distribution--including its little brother, SuSE Linux Professional 8.2--SLES 8 comes in a short two-ring binder that reminded me of the old AT&T System V manuals from decades past. The four CDs are inside the binder, behind their own tab, in a set of plastic sleeves held closed by a one-time-open seal that reads "By breaking the seal you accept the exclusion of warranty." This is an interesting shrink-wrap license; instead of accepting some kind of restrictive EULA, you're accepting the No Warranty clause of the GPL. This is SuSE's way of protecting itself under its native German law, the implications of which are beyond the scope of this article. In short, the laws say that someone has to be liable unless you, the customer, explicitly agree to it.
The manuals themselves (Installation and Administration) are detailed almost to an extreme. The installation manual goes so far as to describe how to set up a VNC client on Windows so you can install the SuSE system from a remote console in a heterogenous environment. The administration manual is detailed similarly. Both manuals have plenty of screenshots, footnotes and everything else a serious 500-server wrangler could want. Perhaps this might be a bit overwhelming for a newbie, but SLES 8 is no newbie's distribution.
Personally, I skipped the manuals the first time and went straight to booting. The installer went straight into framebuffer mode before it ever gave me a boot prompt. Pressing Return got me the usual chicken tracks a kernel load generates, then popped me back into framebuffer-based X, successfully finding the USB mini-mouse I keep around for such an occasion. Without further prompting, YaST2 started and we were off.
The coolest thing about YaST2 in installation mode for a busy network administrator is it is menu-based, rather than dialog based. I simply could have selected Next on the first screen, and it would've started asking me for CDs and pulling packages off of them. Instead, I repartitioned the disk array to my liking, a process that was a little clunky but acceptable. The format utility defaults to ReiserFS; I'm usually an ext3 fan, but for the sake of experimentation I took the default. I also fiddled with the package settings, adding kernel-source and its associated dependencies. YaST2 handled the dependency questions in a pop-up box with a somewhat complicated (but complete with respect to choices) dialog. I told it, yes, I wanted the C compiler, and it even gave me a choice as to which version. The package selector allows you to choose by function (like Debian's tasksel) or by package group (something any RPM-based administrator is familiar with), or you simply can search for a package, an option I haven't seen before in a GUI installer. It worked quite well and quickly.
Clicking on Next back at the main menu screen got me a somewhat different looking progress screen. The system showed me which CDs it would need, in what order and how much time it expected to take on each of them. After feeding it the first CD, it began to update those estimates in real-time. Accounting for switching CDs, the install took about 20 minutes.
The CD prompts are a bit confusing. It asked for CD 1, and I fed it the CD with the big 1 on it, which it refused it until I read the fine print. It wanted United Linux CD 1, which has a big 2 on it and United Linux CD 1 in small print. Once I figured that out, the rest was easy. It did want the first two CDs back, briefly, after it finished with the first one; I'm not sure of the logic behind that.
A reboot brought the machine right back up into X for the postinstall. It asked for a root password, forced me to add a non-root user (it's not optional, I tried), asked if I wanted to change the network settings (it had already figured out I have a DHCP server), asked me about my X configuration and went straight to runlevel 5.
You see, the ULB has a brand-new NVIDIA Quadro card in it, a version that is too new for the default X drivers to recognize. Fortunately, SaX (the X configure tool in SuSE) saved a copy of what it knew was working before overwriting things with my configuration. I simply logged in in console mode as root, copied that back into place and ran startx. Presto, KDE was there with a big red background full of bombs and caution icons. I got the hint; no X as root. So I logged in as myself, and KDE came up again. It starts it with a handy greeter tool that has URLs to register your software, to get into the issues database and other SuSE-specific stuff; the SuSE chameleon head icon on the desktop restarts this part.
The KDE 3.0 desktop was spartan but correctly recognized both the CD-RW and the DVD-ROM and created appropriate icons. The DVDs Play selection on the right-click menu didn't do anything, alas, but it did read the UDF filesystem correctly and let me read the README file on my test disk. Double-clicking the CD-RW icon brought up k2b, KDE's answer to xcdroast and Nero. Other than k2b and the Konqueror and Mozilla (1.0.1) web browsers, though, there wasn't a lot of desktop software. No office suite, no PDA tools not even Emacs in the default setup (though Emacs was available off the CD).
On the other hand, as you would expect from something with Enterprise in the name, the desktop was heavy on servers and server tools. Everything from Apache (with Jakarta-Tomcat) to Zebra was present, far too many to list here. The one thing I did not see was anything resembling an enterprise-class backup tool, not even Amanda, which is GPL. (Of course, correcting that problem should be reasonably easy for a large system administrator, but it would be nice to have something ready to go.)
On the positive side of things is SuSE's swiss army knife configuration tool, YaST2 (Yet Another Setup Tool, Version 2). This thing is nice. It's fast for a GUI, at least on the ULB, and I've watched other GUI tools be slow on dual Athlons and Xeons. Like install mode, YaST2 is menu-driven in double-column fashion. The firewall tool is especially nice; I could walk through the dialog and set up a reasonably effective masquerading firewall in about two minutes and not know the first thing about iptables. I was distressed to discover, in the security tool, that the default mode for password encryption is still DES. Both MD5 and Blowfish are available as alternatives, however, so that helps.
The software installer tool not only already knows what is on all four CDs and which CD it's on, it also incorporates the on-line updater. The hardware detector is thorough in typical German fashion, giving you information on almost everything in /proc in a nice, nested GUI. (Not to worry, hardcore console types; simply leave off the 2 when you type yast and a dialog-based curses interface pops up, same base design only no mouse.)
Despite the little niggly problems I mentioned, I am really impressed with the job SuSE did on the new platform. Usually on something other than pure 32-bit x86, packages are out of date, things don't work quite right and other annoyances abound. SELS 8 has XFree86 4.2, kernel 2.4.19, KDE 3, Samba 2.2--not quite bleeding edge but definitely current. And there are 32-bit libraries (glibc-2.2) on the machine so things you don't have 64-bit packages for can run. And, aside from the NVIDIA driver issue, which is really NVIDIA's issue and not SuSE's, it works--exactly what you want in an enterprise-class OS.
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.
|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
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Linux Systems Administrator
- Introduction to MapReduce with Hadoop on Linux
- RSS Feeds
- New Products
- Weechat, Irssi's Little Brother
- Validate an E-Mail Address with PHP, the Right Way
- Tech Tip: Really Simple HTTP Server with Python
- Poul-Henning Kamp: welcome to
8 min 11 sec ago
- This has already been done
9 min 11 sec ago
- Reply to comment | Linux Journal
54 min 25 sec ago
- Welcome to 1998
1 hour 42 min ago
- notifier shortcomings
2 hours 6 min ago
3 hours 43 min ago
- Android User
3 hours 45 min ago
- Reply to comment | Linux Journal
5 hours 38 min ago
8 hours 27 min ago
- This is a good post. This
13 hours 40 min ago
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?