Letters

Readers sound off.

Letters

KDE Baby

My baby daughter who is only 3.5 months old just loves Linux. After installing SuSE on my new and shiny P4 and introducing her to Geeko the chameleon, she is getting very interested. She especially likes KDE. I guess she is attracted by the Fisher Price Toys-like look and feel of the Keramik theme. The photo shows Idun showing her enthusiasm for KDE.

How many Linux users have children already Linux-enabled? Are babies welcome at Linux Counter?




Janus Sandsgaard

SMART and Disk Errors

In the January 2004 issue of Linux Journal, I came across the article on SMART by Bruce Allen on page 74. I downloaded and installed it without a hitch. I ran it with the -t (test) option and immediately afterward I encountered disk errors. (Before I ran smartctl everything was fine.) I shut down and rebooted. The system ran fsck because it encountered errors on the hard drive. I eventually had to run fsck in manual mode and lost info.


Marshall Lake

Bruce Allen replies: I'm sorry to hear you've had disk problems. What's happened in your case is unfortunate. You had a disk that already was compromised before you examined the SMART data and ran self tests. Probably some of the disk sectors were unreadable, but you didn't know it because they were storing files that you normally don't access. When you ran a disk self test using smartctl -t, this did a read-scan of the disk, and the disk problems became apparent.

Before investing additional time in the system with this disk, make sure that the disk does pass a long self test without errors and that the Current pending sector and/or Offline pending sector raw counts are zero. If not, the disk needs repair/replacement before you fix the OS problems.

If your OS distribution is RPM-based, I suggest that you use the verify options of RPM to make a list of all missing or changed files on your system along with the name of the package that they came from. Then, re-install the packages that have missing or corrupted files.

The failure rate of new hardware (and particularly new disks) is higher than when they are a few days or weeks old. New hardware is more likely to fail than hardware that has worked for some time. The fact that fsck worked okay means only that the filesystem is consistent. It does not read all the data from the disk.

The smartctl -t command starts a self test that reads every byte on the disk. This is a much more thorough test than fsck. I strongly advise you not to trust the disk unless it runs a smartctl -t long test without revealing any problems.

Questions for Robert Love on I/O Scheduler

I just read your I/O scheduler article in the February 2004 Linux Journal. Great article, but it left me with a few questions. Hopefully you can answer them for me. How do the Deadline and Anticipatory I/O Schedulers detect when a deadline has been met or passed? Are there multiple kernel threads, one or more that are processing the main queue and one or more that are walking the read/write queues checking deadlines? Wouldn't the Anticipatory algorithm also give the possibility of causing starvation? If we assume a request comes in and the deadline passes, the I/O scheduler will see that the deadline is passed and process the request, then wait 6 milliseconds thinking there may be one or more requests in the same area. If there are one or more requests in the same area, then will it basically loop processing these close requests, waiting 6 milliseconds, processing more requests, etc.? The test outputs really show how much better the I/O schedulers in 2.6 are than in 2.4 , but how much does the process scheduler play into this—choosing the O(1) scheduler over the default 2.4 scheduler?


Torin Ford

Robert Love replies: Glad you liked the article. There are no threads involved. The I/O scheduler code is handled in two places: in the interrupt handler for the disk's driver and in the process context code of the process submitting the request. The code can detect a deadline easily, because the deadline-sorted lists are inherently FIFO. The request at the head of the list has the oldest expiration, so the list need not be walked—just peek at the head of the FIFO list.

In reply to the question about whether the anticipatory algorithm will also lead to starvation, eventually another request will expire, and it will go service it. The anticipation heuristic is just used to decide where to seek to, it otherwise does not change the algorithm of the disk.

I don't think the better process scheduler comes into play whatsoever. We know why the times are better: the writes-starving-reads phenomenon, so the better numbers are readily explainable.

English Lesson

The February 2004 issue of Linux Journal once again has the awful phrase “comprised of” (on page 56). Can I ask you to grep through incoming manuscripts to catch this annoying solecism in the future?


A. T. Young

It'll be in our stylebook from now on. —Ed.

Flaky Entertainment System

I was on a Virgin plane in December 2003 with such a faulty system that something like 90% of us mostly saw the reboot console! Luckily, it was a night flight, otherwise the kids would have gone crazy, and there was no mention of Linux on the screen, so most people would not have associated it with our beloved OS! However, it also shows that we need to release only well-tested stuff to real customers, a lesson we need to learn if Linux is to move into the home environment.


Kirk Martinez
University of Southampton

Compatibility List?

Recently I went looking for a Linux compatability list on the Web and I was unable to locate one. Does LJ know where one is located?


Walt L. Williams

Every major Linux distribution maintains a hardware compatibility list. Not all distributions include the same hardware support, so you need to check yours. If you want to know whether there's a driver available in source code form that you can compile and configure, you usually can do best by searching for the hardware name +linux on Google. —Ed.

Table Fix for Work-Flow Article

I noticed a mistake in the February 2004 issue of LJ in the article by myself and Luciano Barone about REDACLE. In the text there is a reference to Table 2 that is missing. Table 2 should, in fact, read as follows:

IDNameSubnameType
195crystalBarrel1L
36capsuleBarrelC1
35apdBarrel 

Only the first row is relevant for text comprehension. Tables actually marked as 2 through 6 should be renamed 3 to 7.


Giovanni Organtini

Photo of the Month

Here are a couple of photos of our last vacation. We went to Roswell, New Mexico to find out if the aliens were using Linux. Most were Debian supporters. There is a before and after photo. Some help from The GIMP was required.




Tom and Orion Maier

Photo of the month gets you a one-year subscription. Send to ljeditor@linuxjournal.com. —Ed.

Lindows: Good install, but Support Needs Work

After reading Steve Hastings' review of Lindows 4.0 [LJ, December 2003], I had a few comments of my own to add. First, concerning the Click-n-Run warehouse, I had many of the same complaints as Steve. Along with the rebranding of many packages and the inability to get packages to work properly, there is a serious problem with software versions being behind the curve. For instance, their OpenOffice.org version is 1.0.3, whereas the current OpenOffice.org is 1.1.x. Evolution in the warehouse is on version 1.2, while Ximian is on version 1.4. There are many examples of this.

Another serious problem is in the area of customer support. I sent a report to Lindows tech support in late December 2003 about these issues and have yet to receive anything other than an automated response (“We'll reply to this issue in a few days”). Since that initial report, I've sent two follow-ups, and those have gone unanswered as well.

Having said all that, I will agree that Lindows is very simple to install. I tried both 4.0 and the new 4.5 version, and they install in less than 15 minutes. There were problems with the video card (an ATI All-in-Wonder) and the media-card reader, but I have less problems with getting Lindows to work with the hardware than I did with Mandrake, Red Hat and that other company from Redmond.

But, until they have current version applications available in the Click-n-Run warehouse and until they resolve the issues with customer support, I won't be recommending Lindows to anyone.


Ed Dulaney

Big News at Linux Vendors?

The Linux magazines have been conspicuously quiet about two of the biggest Linux stories of 2003. The first is Red Hat rolling their Linux into Fedora Core and dropping support of their older versions. The second is Novell buying SuSE and Ximian. Red Hat is putting a positive spin on their story, while many others believe it's a blunder on Red Hat's part and a betrayal of their customers. Novell, is now a major Linux player. Does it all of a sudden have Linux religion? Will they contribute to Linux's evolution? Or is Linux just the latest strategy/weapon in their age-old war with Microsoft? Both these stories have a big impact on the Linux community, and they are getting a lot of coverage and discussion on-line. I would expect LJ to join in on the same. Will this happen soon?


Henry E. Alubowicz

See the next letter for why these stories might be less important than you think. —Ed.

LJ Discovers IT's “Dark Matter”

Just wanted to commend Doc on a great piece (“DIY-IT...” [LJ, February 2004]) and LJ for being a brave enough to print it, despite its vendor-specific (read: advertiser) implications.

Doc is on to the scoop of his journalistic career with his series on how open source (and LAMP in particular) is taking over the enterprise and putting customers back in control. The game is no longer about getting spoon-fed solutions that take too long to implement, cost too much money and fail to address the specific needs of any given business. It's time to say good-bye to cookie-cutter enterprise software and Doc knows it.

What is more, another reason why he is on to something big is because the open-source way, and the LAMP platform in particular, make developers more productive. Vendors know it, developers who have used the tools know it, and even those that haven't are beginning to suspect it.

In my day-to-day, I deal with friends and colleagues (as well as my own staff) whose employers everyone would recognize as being pillars of the Internet economy. These folks depend daily on the huge productivity improvements that come from open development tools and the passionate communities willing to stand behind them. However, most of them either won't, or can't for various corporate reasons, climb the nearest tree and scream to the world about it.

Like dark matter that can't be seen but whose presence is felt everywhere, all of these folks “in the know” are rebuilding the very infrastructure on which IT depends, one brick at a time. And, in perhaps what is just as important, they are also rebuilding all of the starting assumptions behind what “enterprise software” is all about. So thanks Doc and LJ. Keep on it—you're on to something big.


Antonio Rodriguez
VP Engineering

Errata

In the article “DIY-IT: How Linux and Open Source Are Bringing Do-It-Yourself to Information Technology” [LJ, February 2004], Craig McLane, VP of Technology at Ticketmaster, was mistakenly identified as the source of quotations that were in fact made by Sean Moriarty, EVP of Products and Technology at Ticketmaster. The article also said the event where the talk took place was LinuxWorld Expo. In fact, both men spoke at the O'Reilly Open Source Convention. —Doc Searls

______________________

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Photo of the Month Tom Maier

Dave McHenry's picture

Hi,
Trying to reach Tom and Orion Maier who posted picture of the month.
Can you help me?

Webinar
One Click, Universal Protection: Implementing Centralized Security Policies on Linux Systems

As Linux continues to play an ever increasing role in corporate data centers and institutions, ensuring the integrity and protection of these systems must be a priority. With 60% of the world's websites and an increasing share of organization's mission-critical workloads running on Linux, failing to stop malware and other advanced threats on Linux can increasingly impact an organization's reputation and bottom line.

Learn More

Sponsored by Bit9

Webinar
Linux Backup and Recovery Webinar

Most companies incorporate backup procedures for critical data, which can be restored quickly if a loss occurs. However, fewer companies are prepared for catastrophic system failures, in which they lose all data, the entire operating system, applications, settings, patches and more, reducing their system(s) to “bare metal.” After all, before data can be restored to a system, there must be a system to restore it to.

In this one hour webinar, learn how to enhance your existing backup strategies for better disaster recovery preparedness using Storix System Backup Administrator (SBAdmin), a highly flexible bare-metal recovery solution for UNIX and Linux systems.

Learn More

Sponsored by Storix