Best of Technical Support
I have a dual-OS machine running Microsoft Windows XP and SuSE 8.2.
When I do a shutdown in XP my PC powers off. But when I turn off
this computer in Linux, my PC does a reboot.
This is most likely due to inherent problems between SMP and APM. The
two standards are mutually incompatible, apparently resulting from an
unavoidable race condition among the multiple processors.
In Linux's case, APM is disabled in SMP kernels, even if those happen to
be running on single-processor machines. You might try switching to
the UP (uni-processor) kernel, or you could compile your kernel with an
option that forces the APM power-off feature to work.
Try passing apm=power-off to the kernel at boot time.
Usman S. Ansari
I'm having a problem configuring my sound card for my Red Hat 7.2 system. When I run sndconfig, it comes back that I have an ISA PNP card, a SoundBlaster 32 Wavetable card to be exact. When it goes to test the sound I get these errors:
/sb.o : init_module: no such device /sb.o : insmod /sb.o : failed /sb.o : insmod sound-slot-0 failed
When I run dmesg I get this:
sb: No IsaPnP cards found, trying standard ones... sb: I/O, IRQ, and DMA are mandatory No detected device
I get that same message when I try any different SoundBlaster card
in the list.
It sounds like your card is configured with I/O base address, IRQ (interrupt request) and DMA (direct memory access) settings that the kernel can't autodetect. You might have to add a line to your /etc/modules.conf that looks something like:
option sb io=0x220 irq=5 dma=1 dma16=5 mpu_io=0x330
where you replace the numbers with those that your card is using.
My broadband ADSL Internet connection has a monthly fixed fee
if downloads don't exceed some maximum limit; at the time of this
writing, the limit is 3GB. I'm using PPP over Ethernet to connect to
the Internet. I want to know if there is some application that can
let me have an account of the transferred bytes over my connection.
Guillermo Gimenez de Castro
There are several. ipac is the IP accounting package. MRTG is the
multirouter traffic grapher, which may be overkill for your needs,
as it graphs usage rather than simply totaling it. You also can use
the ifconfig command and look at the received (RX) and transmitted
All in all, ipac is probably the simplest package for you to
examine. See www.daneben.de/ipac.html.
Also, the sar command can provide you with several statistics about your network interfaces; as root try the command:
sar -n FULL
which will provide you with transmitted/received packets and bytes among
other information in a timed table. Do a man sar for further information.
Felipe Barousse Boué
I am trying to install Slackware 9.1 on a new system that uses a
serial ATA drive. The system uses an Intel motherboard with two
two-device IDE ports and two serial ATA ports. I can have a total of six
devices set up with this rig. The two IDE ports have two CD-ROM
drives and a Zip drive attached to them. One SATA port has the only
hard drive, the other is unused.
I am booting from Slack's installation CD, and the boot proceeds
normally until it gets to the point where the drives are discovered.
The system sees all of the ports and sees the SATA drive as hde on
IDE2. I have partitioned it so that hde4 is the partition that I
want Slack to recognize. The system knows the drive is there, but it
stalls at a point with this message hde4: loading IDE
drivers. I can't
get any further than this.
What should I look for in dealing with SATA drives on Linux?
You will find yourself doing
some tweaks depending on the hardware and distribution you have. A good
starting point to find out more about this is
you will find tips specially related to incompatibilities of
controllers, drives and Linux.
After I picked up my first copy of Linux Journal on
the newsstand, I
couldn't believe that I'd found in one issue the answers to several
of the problems I've been up against. I immediately bought that
issue and subscribed on-line as soon as I got home.
After digesting the information on Nagios, I'm now looking to replace
my company's current DLT backup solution. Our current environment
runs a Windows/Veritas Backup Exec 8.6 solution and we're paying
a hefty price for these systems. My question is this; is there a
viable Linux solution that supports a wide variety of tape backup
hardware? Currently I have four single DLT 15/30GB drives at the
office, but I also need to support a seven-tape DLT autoloading library
on my home LAN. I'd like something that doesn't require a huge
investment in time to learn. After all, it only takes the average
user about two clicks of the mouse to lose a file, so I'd like to
be able to restore it as easily.
BRU (Backup and Recovery Utility) is reasonably well regarded. It's
proprietary but not very expensive. More information can be found
BURT (BackUp and Recovery Tool) is at the University of Wisconsin,
www.cs.wisc.edu/~jmelski/burt, just as AMANDA was created at
the University of Maryland.
Backup software isn't necessarily the best reason to choose Linux,
not because it isn't available, but because it's often the same product.
Most of the major commercial vendors of backup solutions
now support Linux. There are also mid-range solutions that
are more cost-effective but still provide graphical wizards and management
If you would prefer an open-source solution, there is a wide variety of
these options available as well, but you also could simply rely on good old
tar and gzip or something more robust, such as cpio. You will need the
magnetic tape tools package, mt, and the appropriate driver(s) installed in
your kernel. If you do go with a tape library, you also may need to search
around for a utility that controls the media loader on the device, so you may
want to do some research ahead of time before you buy one.
We have found that using our own scripts (mostly in Python) for
backups (local and distributed), backup verification and validation and
restores has been the best alternative so far for the different backup
needs we have. We do perform backups into tape devices, CD-based
technologies and into other physical hard disks as data and disk backup.
A couple of references: www.linux-backup.net has various pieces of information
regarding backups in Linux; also look at
the book Unix Backup and Recovery, which
Linux Journal reviewed a while ago. Although the book
is a bit old, it may still be worth
reading. The LJ review is at
On the hardware side, check the site www.linuxtapecert.org.
Felipe Barousse Boué
I've been using UNIX in some form or other for over two decades.
Using Red Hat, I put out camera-ready copy for my latest book,
The Economy and Material Culture of Russia,
larger format, with 104 graphs produced by Stata from 108,000
records in filePro16. My exceptionally handy word processor was
WordPerfect for the camera-ready copy.
Now the university is forcing me to upgrade my computer, which will
have Red Hat Linux 9 on it. My understanding is that Corel no longer
maintains WordPerfect, which won't run on Red Hat Linux 9. What
is the most suitable word processing package for this project?
What do you recommend?
There are many word processors available, and your choice of them
depends on your publishing needs. You should begin by examining
the ever-present Emacs and the LaTeX and SGML document description
languages. Most people find that these are too obfuscated to suit
their needs, but it's always worth the examination as these are
extremely powerful document layout products once you know how to
If you prefer a WYSIWYG word processor, you can install
OpenOffice.org or KWrite, both of which are open-source products. Or,
if you need better compatibility with Microsoft Office users, you
can try either Sun's StarOffice product, which is OpenOffice.org
with additional fonts and commercial support, among other things,
or IBM's Lotus SmartSuite, which is also a commercial product.
These are only a few of the options available, and these options
do not even include the desktop publishing products. Take a look
around—you might be surprised at the variety of options available.
Like much unsupported proprietary software, you can keep WordPerfect
going by installing old versions of libraries (linuxmafia.com/wpfaq).
If you want to keep the ability to import Microsoft Word documents,
you need to apply another fix, too: www.linuxjournal.com/article/5655.
I know I can't be the only one to exhibit this embarrassing
behavior on a semiregular basis, so here goes. For whatever reason,
there have been times where I inadvertently disclosed a sensitive
password on the command line, mistakingly thinking that my input was going
to the stdin of a different program such as ssh or smbclient. I use the
bash shell, so this means my carelessness gets written to a history
file. Normally, this isn't too big of a problem, but sometimes I end up
using a shared account on the system. Needless to say, whoever else
has access to this account ends up being able to view my password
in the history file. Is there an easy way of telling bash to discard
entering a prior or specific entry into its history? I'd rather
not have to edit the history file manually, which seems to be the
only way I know to cover my tracks.
First, if you realize your mistake before you press Enter, simply
press Ctrl-U. Doing this erases all of your typing on the current line.
This works at the shell prompt (most Bourne-compatible shells), the
login prompt and even in vi (while still in insert mode).
If you've already pressed Enter, then your fastest, easiest recourse is
re-read the history file that's already on the disk. Since the
history normally is written only on logout, this will overwrite the in-memory history. Type history -r ~/.bash_history.
Of course, this also will wipe all of the other entries from the current
session, and it will be as if you just logged in (as far as your history goes).
Do a man history to check on the options of the history command:
history [n] history -c history -d offset
With no options, history displays the command history list with line
numbers. A numerical argument of n lists only the last n lines. The
-c or -d options, if supplied, have the following meanings: -c, clear
the history list by deleting all the entries and -d, offset Delete the
history entry at position offset.
Felipe Barousse Boué
Is there a Debian-based distribution that would allow me to
install on an SATA hard drive? The hard drive controller is a
Micro-Star International RAID Bus Controller, using ata_via per
hardware identification by YaST, and/or a VIA 8237 per the MSI
KT6 Delta mainboard manual. I want to switch back to Debian, but
the installer does not recognize that my system has a hard drive.
The same goes for using Knoppix's knx-hdinstall. I understand that
I can install a live system on an IDE hard drive, add the modules
necessary to get SATA support working, copy the whole thing over to
the SATA drive and then run LILO to get the system working on the
SATA drive, but that sounds a bit too complicated for someone as
lazy as I am.
Well, you can install Debian on just about anything if you bypass
its normal installer and use the debootstrap package. There are
some tricks to using that, however.
On my Wiki pages I have described a technique for installing
Debian onto a set of disks under LVM (logical volume management), using nothing but an LNX-BBC (www.lnx-bbc.org) and my
So, if you have a rescue disk like the LNX-BBC that can see and
access the SATA hard drives, you could follow basically the
same procedure that I describe on my pages (www.starshine.org/sysadmoin/DebootstrapInstallation).
I will warn that this is not easy. It is somewhat laborious and
my step-by-step description doesn't go into much explanation.
It assumes expertise in partitioning (using fdisk), making and
Xandros Desktop OS (xandros.com), LindowsOS (lindows.com) and
Libranet GNU/Linux (libranet.com) are all Debian-based and maintain
hardware compatibility lists. You can look up your Serial ATA
hardware on their Web sites.
Webinar: 8 Signs You’re Beyond Cron
11am CDT, April 29th
Join Linux Journal and Pat Cameron, Director of Automation Technology at HelpSystems, as they discuss the eight primary advantages of moving beyond cron job scheduling. In this webinar, you’ll learn about integrating cron with an enterprise scheduler.Join us!
- DevOps: Better Than the Sum of Its Parts
- Return of the Mac
- Drupageddon: SQL Injection, Database Abstraction and Hundreds of Thousands of Web Sites
- Play for Me, Jarvis
- Non-Linux FOSS: .NET?
- Not So Dynamic Updates
- Designing Foils with XFLR5
- Users, Permissions and Multitenant Sites
- April 2015 Issue of Linux Journal: High-Performance Computing