Linux-Hardware Digest #5, Volume #13 Thu, 8 Jun 00 12:13:08 EDT
Contents:
Re: S3 Savage 2000, ATI Rage 128GL or Matrox G400 (Rodd Snook)
Re: slow SCSI performance on Dell Poweredge 6300 ([EMAIL PROTECTED])
4.3GB hard disk on old ide interface ("Martin Peck")
Re: modules on separate /usr - can't initialize from loadlin ("Gene Heskett")
RH 6.2 on Cpq Proliant 5000 - install OK but no boot (Dick Repasky)
Re: Heads, et al...Confused in NF, Canada...*smile* (Bert Sainz)
Motherboard Supermicro PIIIDM3 ([EMAIL PROTECTED])
Re: PCI sound cards (Dances With Crows)
Re: ASUS K7V KX133 motherboard problems ([EMAIL PROTECTED])
Re: Token Ring Adapter (=?iso-8859-1?Q?S=E9bastien?= Cottalorda)
parallel port CD-RW (Helder Correia)
Re: Turning on joystick port (John Gluck)
Re: h/w spec for linux oracle app server - HELP! (John Gluck)
[Help:] Advice on SMP machines required (Jose Manuel Benitez Sanchez)
----------------------------------------------------------------------------
Subject: Re: S3 Savage 2000, ATI Rage 128GL or Matrox G400
From: [EMAIL PROTECTED] (Rodd Snook)
Date: Thu, 08 Jun 2000 12:11:37 GMT
[EMAIL PROTECTED] (Sasa Janiska) wrote in
<[EMAIL PROTECTED]>:
>I need advice which graphic card to pick for use on Linux.
>
>I don't need card for games, its use is primarily for text processing
>and I need good video output for DVD.
>
>I cannot decide between: Diamond Viper II Savage 2000 (32MB), ATI All in
>Wonder 128GL (32MB) and Matrox G400 AGP Dual Head (32MB).
>
>Which card has good support for Linux (XFree) and good quality of video
>output?
>Sincerely,
>Sasa
For quality output and Linux support, every review I have ever read
recommends Matrox cards. Independent (games site) reviews of the G400 all
praise the output quality of the card. I use my display system mostly for
text processing and 2-D gaming. I have a G200 and find it to be an
excellent device for this purpose.
Rodd
------------------------------
Date: Thu, 08 Jun 2000 14:42:30 +0200
From: [EMAIL PROTECTED]
Crossposted-To: comp.periphs.scsi,linux.dev.scsi,ch.comp.os.linux
Subject: Re: slow SCSI performance on Dell Poweredge 6300
Marc SCHAEFER wrote:
>
> In ch.comp.os.linux [EMAIL PROTECTED] wrote:
> > The test was run with no load on the system and only one user.
> > I find these results surprisingly poor; so if anyone has an idea
>
> Yes, they are.
>
> What does hdparm -T -t /dev/sda say ?
# hdparm -T -t /dev/sdc1
/dev/sdc1:
Timing buffer-cache reads: 128 MB in 0.94 seconds =136.17 MB/sec
Timing buffered disk reads: 64 MB in 3.31 seconds = 19.34 MB/sec
> What about creating the fs -b 4096 and re-running the benchmark ?
The blocksize is already 4k.
(ext2 filesystems are by default created with 4k blocksize under
RH 6.2 for 18GB disks).
> How much RAM reported by free ? What does vmstat 5 say when
> run as the same time as Bonnie (mostly read, mostly write, many
> swapping ?)
There is 256MB or RAM, the system does not swap when
running bonnie.
> What version of the kernel (uname -a) ? Red Hat 6.2 is not
> exactly a kernel version (and I don't use Red Hat).
kernel is 2.2.14-5.0
> I get routinely around 17-18 MByte/s with Seagate Chetah and
> AIC7xxx. Although Adaptec controllers are not that fast compared
> e.g. to Symbios Logic (Tekram: better, cheaper).
>
> Also, I got upto 70 MByte/s in a RAID0 configuration (45 MByte/s
> on a fast machine with RAID5).
I am also going to use software RAID, but prior to that I'd like
to make sure I get the best performance (actually, I noticed the
slowness because it took 70 minutes to reconstruct a 18GB disk
in RAID1).
=====================================================================
Dr Loris Renggli, BSP, Universite de Lausanne, CH-1015 Lausanne
IT manager e-mail: [EMAIL PROTECTED]
tel: +41-21-692 3603 fax: +41-21-692 3605
------------------------------
From: "Martin Peck" <[EMAIL PROTECTED]>
Subject: 4.3GB hard disk on old ide interface
Date: Thu, 8 Jun 2000 13:52:49 +0100
Reply-To: "Martin Peck" <[EMAIL PROTECTED]>
hi
I have spent a considerable amount of time reading up about this and trying
out everything I can, but can't quite get this working. Apologies for quite
a long post; any help would be very much appreciated.
I have a 4.3 GB disk I'm trying to put in to a 486/33 for use as a gateway
and webserver. The system works fine with the current 200MB disk. I'm
running Debian linux (kernel version 2.2.15) (the 'potato' version - most
recent stable version).
I can boot from floppies using a kernel with the Enhanced IDE driver. This
properly detects the drive's geometry (9042/15/63). The installation often
proceeds fine; I can partition the disk in to a small /boot partition, a ~
600MB / partition and the rest I partition for /var. These partitions then
initialise and mount without any problems. The installation usually
proceeds ok for a while after that, but at some point it usually runs in to
problems, with, for example, an error message such as:
Kernel panic: EXT2_FS panic (device ide0(3,2)): load_block_bitmap:
block_group >= groups_count - block_group = 131071, groups_count = 5
Sometimes, on the other hand, initial installation goes ok... I then boot
from floppies in order to insert the 'linear' keyword in to the lilo
configuration (without which it doesn't get beyond 'LI' when booting from
the hard disk), and hda4 fails to mount as /var cleanly - masses of error
lines of the form
<2> EXT2_FS error (device ide0(3,4)): ext2_check_blocks_bitmap: Block #n of
the inode table in group 27 is marked free
where in the errors which can be seen on the screen n runs up to 48. [When
initialising hda4, it has the following configuration: 28 block groups,
32768 blocks per group, 32768 frags per group, 16288 inodes per group, block
size 4096, frag size 4096, 456064 inodes, 911688 blocks, 45584 blocks res
for super user.]
I've tried using boot floppies with a kernel with the 'old hard disk' driver
compiled in instead. I've tried parsing hda=cyl,head,sect and
ide0=0x1f0,0x3f6,14 arguments to the kernel, and during the initial boot
sequence the driver seems to see the disk ok - it detectst that it has 4
partitions, for example - but the installation system subsequently can't
detect that a hard disk is present.
With the larger default boot floppy kernel, the same things happen as when
comiling in just the newer EIDE driver.
The disk itself works fine with a newer PII computer so that's not the
problem - and the computer works fine with a smaller disk as I said above.
The BIOS of the 486 even allows me to enter the full geometry of the large
disk.
Any ideas on what else I could try would be very gratefully received.
thanks,
Martin Peck.
------------------------------
Date: 8 Jun 2000 8:52:59 -0500
From: "Gene Heskett" <[EMAIL PROTECTED]>
Subject: Re: modules on separate /usr - can't initialize from loadlin
Unrot13 this;
Reply to: <[EMAIL PROTECTED]>
Gene Heskett sends Greetings to Yuri Oskotsky;
YO> Sorry for the rushed post, I have already figured this one out:
YO> autoboot folder contains kernel 2.2.5, but the kernel installed
YO> at installation time is 2.2.19. And therefore dependencies were
There isn't yet a 2.2.19 kernel. 2.2.16 was just published yesterday.
Cheers, Gene
--
Gene Heskett, CET, UHK |Amiga A2k Zeus040, Linux @ 400mhz
Ch. Eng. @ WDTV-5 |This Space for rent
RC5-Moo! 350kkeys/sec, Seti@home 16 hrs a block
email gene underscore heskett at iolinc dot net
This messages reply content, but not any previously quoted material, is
� 2000 by Gene Heskett, all rights reserved.
--
------------------------------
From: [EMAIL PROTECTED] (Dick Repasky)
Subject: RH 6.2 on Cpq Proliant 5000 - install OK but no boot
Date: 8 Jun 2000 13:41:25 GMT
Reply-To: [EMAIL PROTECTED]
- Compaq Proliant 5000 with EISA raid controller card.
This machine is not in Compaq's list of RH certified machines
- Red Hat 6.2 installs ok but won't boot afterwards
'partition table type not recognized'
kernel then panics because it cannot mount root fs
- This is repeatable over many installs.
- I believe that the problem is that the partition table format used
in the installer kernel is not in the installed kernel. In the
source code, the error message is issued after all partition table
types have been attempted. In each attempt, the partition table is read
from the disk. If that fails, an "unable to read partition table"
message is issued. Hence, I believe that the kernel is successfully
reading from the disk but doesn't recognize the table.
- Forcing a DOS partition table using Compaq's disk configuration
tool did not work. (Told tool that sys to be installed was NT.)
- We are thinking of trying the following. Bring up the installer
into rescue mode, mount the file systems, chroot into installed
root file system, configure, compile and install a kernel
that contains all partition table types. Reboot.
Has anyone had this problem before?
Does anyone know what is wrong, how to fix it?
Does our final effort seem reasonable?
Thanks,
Dick
---
------------------------------
From: Bert Sainz <[EMAIL PROTECTED]>
Crossposted-To:
comp.sys.ibm.pc.hardware.storage,alt.comp.hardware,alt.comp.hardware.homebuilt,alt.comp.hardware.homedesigned,alt.comp.hardware.pc-homebuilt
Subject: Re: Heads, et al...Confused in NF, Canada...*smile*
Date: Thu, 08 Jun 2000 10:01:03 -0400
Hendrix wrote:
> Hendrix wrote:
> >
> > Hi guys,
> >
> > Until recently, I understood two meanings of the word "head" in
> > conjunction with a hard drive... One understanding of mine was that the
> > word "head" referred to the read/write head that is used to read/write
> > data from a particular side of a platter. Another understanding I had
> > was that a single side of a platter was referred to as a head. Thus, a
> > hard drive with 4 platters would have 8 heads (ex: 4 platters x 2
> > sides/platter = 8). Whether you're counting read/write heads, or sides
> > of a platter, you will still come up with 8 heads...
> >
> > Does this sound right...??? It sounds right to me...EXCEPT...I have a
> > hard drive here that I know has only 2 platters, but yet it is rated as
> > having 16 heads... It is a 2.57GB Fujitsu that is rated as having a
> > geometry of 4982Cyl 16Hds 63Sec/Trk....!!! And if my previous
> > calculations are correct, the head count would yield 8 platters...
> >
> > Also, I have another misunderstanding regarding this calculating
> > procedure... For instance, if the number of cylinders on a drive is
> > equal to the number of tracks on each side of a platter then (Cylinders
> > x (Number of Sides) would equal the number of tracks that are on the
> > disk... Now, if I have the number of tracks on a disk, then all I would
> > have to do to calculate the size of the drive would be to take the
> > sectors/track information, multiple it by 512, and then multiple it
> > again by the number of tracks on the disk... Logically this should
> > yield the amount of space on the drive. But it doesn't...
> >
> > Consider the following:
> > (Fujitsu 2.57GB with 2 platters)
> >
> > C=4982 H=16 S=63 BPS=512
> > SizeOfDrive = C x H x S x BPS
> > = 4982 x 16 x 63 x 512
> > = 2571190272
> >
> > This is correct... But consider the following (which would logically be
> > the same):
> > (Still using the 2.57GB Fujitsu with 2 platters)
> >
> > C=4982 H=16 S=63 BPS=512
> > NumberOfTracks = C x NumberOfPlatters x 2 (There are 2 sides to every
> > platter)
> > = 4982 x 4
> > = 19928
> > SizeOfDrive = NumberOfTracks x S x BPS
> > = 19928 x 63 x 512
> > = 642797568
> >
> > So my question is... Why are these calculations coming out differently
> > if their respective meanings and translations are the same
> > throughout... The only thing I can think of is that I am
> > misunderstanding the meaning of a head... If a head is supposed to
> > represent a platter surface, or a magnetic reading and writing device,
> > then there is no way in hades that there are 16 heads on a 2 platter
> > drive...*smile*
> >
> >
> The confusion still continues with the older drive that I have here...
>
> I have a 177MB drive with the following specs C = 903 H = 8 S = 46...
> Why is it saying that there are 8 heads when I know there are only 2
> platters....??? Logical assignments shouldn't even come into play when
> the drive was first made in the early 1990's with 177MB of storage...
>
> It should be easy to convert cylinders to tracks just by multiplying
> cylinders by the number of surfaces on the drive... But this doesn't
> work.... Hmmm...
That drive is actually a C=1806 H = 4 S=46 or some combination of numbers
that will multiply out to the same value of 332,304 sectors. There is a
limitation on these numbers. The maximum number that DOS can use for C is 1024
(or 1023, I forget), for H it is 16 and for S it is 63. So, to end up with the
correct number of sectors you lie to the BIOS which in turn lies to the OS and
you end up with the correct number of sectors. The OS will use an algorithm
that gives the sector number when accessing the drive and does not care if it
is being lied to as long as it can get the right sector. And also, the drive
electronics knows the lie and understands what needs to be done when it gets
the information to read a sector.
So, when you see the term "logical" replace with "damn lie" and you might
understand what is going on. There are two concepts and the one you are
thinking of is "physical" instead of "damn lie"... oops... "logical".
Bert
------------------------------
From: [EMAIL PROTECTED]
Subject: Motherboard Supermicro PIIIDM3
Date: 8 Jun 2000 13:54:53 GMT
Does anybody have any expireence with Supermicro PIIIDM3 motherboard
with 2 processors installed?
For some reason it functions perfectly when there is one processor in
it. But, when two processors installed it boots from smp (but fails with
non-smp) 2.2.15 kernel on raw floppy (cat /vmlinuz>/dev/fd0) but fails
with either lilo (ver 21) or syslinux.
With non-smp kernel it hangs on SCSI initialization.
With smp kernel and lilo it hangs just on unpacking linux.
Is there something which could be done (i.e. bios upgrade)
safe just returning motherboard to dealer?
==================================================
Victor Wagner [EMAIL PROTECTED]
Programmer Office:7-(095)-785-09-72
Communiware.Net Home: 7-(095)-135-46-61
http://www.communiware.net http://www.ice.ru/~vitus
------------------------------
From: [EMAIL PROTECTED] (Dances With Crows)
Subject: Re: PCI sound cards
Date: 08 Jun 2000 10:14:25 EDT
Reply-To: [EMAIL PROTECTED]
On 8 Jun 2000 02:25:40 GMT, Munge
<<8hn074$2tm$[EMAIL PROTECTED]>> shouted forth into the ether:
>Can anyone recommend a supported PCI soundcard for Linux ?
>Preferably an inexpensive model, I'd use ISA but I've run
>out of ISA slots.
Ensoniq AudioPCI97... $25-30 US, available lots of places, install under
Linux is as simple as "modprobe es1371" or "modprobe es1370". Sound
quality isn't the greatest, but you won't notice unless you're an
audiophile, and if you were one, why would you be looking for an
inexpensive sound- anything?
--
Matt G / Dances With Crows \###| You have me mixed up with more
There is no Darkness in Eternity \##| creative ways of being stupid?
But only Light too dim for us to see \#| Beer is a vegetable. WinNT
(Unless, of course, you're working with NT)\| is the study of cool. --MegaHAL
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: ASUS K7V KX133 motherboard problems
Date: Thu, 08 Jun 2000 14:18:04 GMT
Hey! I got the same crap over here. My k7v just came in yerterday. Set
it all up. Took my old Linux drive off my old machine and installed it
in the k7v. Booted off the CD (RedHat 6.2) because it wont boot of it.
Installation starts, and as soon it hits hda for checking... (BOOOOOM!)
I/O errors galore and resets all over the place. Installation fails.
What's up with that????
System:
ASUS K7V
128mb Corsair PC133
AMD Athlon 800mhz
Matrox G400 32MB Dual Head
In article <[EMAIL PROTECTED]>,
Ryan Sackenheim <[EMAIL PROTECTED]> wrote:
> I just recently bought an ASUS K7V motherboard, and I've had problems
with
> hard drive accessing in Linux. First of all I moved my hard drive
over
> from my old system, and had some major filesystem errors, in which I
ended
> up having to completely wipe the drive clean. And I have not yet been
able
> to get any Linux OS installed on the machine. I have tried both
Mandrake
> 7.0 and Red hat 6.2, and they both either lockup or reset after
attempting
> to format or install.
>
> The strange thing is that it appears to work for a period of time,
but
> then drops out with a kernel panic. I have attempted to change some
of the
> settings in the BIOS to disable the Ultra/66 mode for the hard drive,
and I
> even moved it to the slower controller, but I still keep having the
same
> errors. I don't think it is a hard drive problem, because the hard
drive
> is about 6 months old, and was previously working just fine. Also I
was
> able to install Win98 without any problems, so I don't believe it is a
> hardware problem. I was able to use a floppy distribution (Tomsrtbt),
and
> mount the drives without any problem, as well as copying files. But
like I
> said, the problem appears to happen at random (except during
installation).
>
> Has anyone had similar experiences using ASUS's K7V board? I can't
seem
> to find much about it's compatibility with Linux. Is there a way to
get
> Linux to work with this board? Here's some general info about my
system.
>
> ASUS K7V KX133 Athlon motherboard
> AMD Athlon 650
> 128MB 133x RAM
> 13GB IBM HDD
>
> [EMAIL PROTECTED]
>
> --
> Posted via CNET Help.com
> http://www.help.com/
>
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: =?iso-8859-1?Q?S=E9bastien?= Cottalorda
Subject: Re: Token Ring Adapter
Date: Thu, 08 Jun 2000 16:54:38 +0200
Hi,
Thanks for your answer, but I read that only "ISA" IBM Token Ring card
are supported.
Sebastien
Tobi wrote:
> S�bastien Cottalorda wrote:
> >
> > Hi all,
> >
> > I need to buy a Token ring adapter, but I don't know wich one can work
> > on my config :
> > Linux RedHat 6.0
> >
> > the cards :
> > - Olicom RapidFire 3140 (Token Ring PCI 16/4 Adapter)
> > - Olicom RapidFire 3141
> > - IBM Token Ring Adapter PCI 16/4
> >
> > Or if you have any other card to suggest ...
> >
> > Thanks in advance
> >
> > Sebastien
>
> Install the kernel sources and run "make menuconfig" or "make xconfig".
> Then look at the token ring adapter section to see which devices
> are supportet. Me thinx the IBM is in there.
> But with most cards you buy, theres a linux driver on disk.
>
> tobias
------------------------------
From: Helder Correia <[EMAIL PROTECTED]>
Subject: parallel port CD-RW
Date: Thu, 08 Jun 2000 16:05:36 +0100
I have an HP 7500 series CD recorder (exsternal) 2 x.
I would like to know whose kernel modules I have to prove, the
parameters, etc...
I have a RedHat 6.2 system.
Thank you!
------------------------------
From: John Gluck <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.redhat
Subject: Re: Turning on joystick port
Date: Thu, 08 Jun 2000 11:03:24 -0400
bernie wrote:
> I have an intel motherboard with onboard sound (creative es1373) and an
> onboard joystick port. I have installed the joystick RPM package. When
> I run the jstest command it tells me that /dev/js0 cannot be found.
> Also, when I type insmod joy-sidewinder it tells me "Device or resource
> busy". The documentation tells me that I may have to turn on my
> joystick port. How to I go abouts doing this with an onboard joystick
> port? Is the joystick port still related to the sound chip even though
> it is on board? I have sound enabled.
>
> Thanks for any help
> bernie
Do an ls on /dev/js0 if it doesn't exist you will need to make the device
node.
If you have installed the source tree for the kernel, there is a file
called devices.txt in the documentation directory. This lists all the major
and
minor numbers for various devices.
"man mknod" will give you the info on how to make the device node. It goes
something like "mknod </dev/whatever> <major> <minor>
<type>".
You may have to change the permissions to make it accessible by everyone.
--
John Gluck (Passport Kernel Design Group)
(613) 765-8392 ESN 395-8392
Unless otherwise stated, any opinions expressed here are strictly my own
and do not reflect any official position of Nortel Networks.
------------------------------
From: John Gluck <[EMAIL PROTECTED]>
Subject: Re: h/w spec for linux oracle app server - HELP!
Date: Thu, 08 Jun 2000 11:08:22 -0400
[EMAIL PROTECTED] wrote:
> Hi,
>
> I need to set up an Oracle application server on an Intel based linux
> server (probably Redhat). The server will have 10 users. Can anyone
> tell me what sort of harware spec I should be looking at for such a
> system? specs such as CPU speed, RAM, and HD space would be excellent.
>
> Any help much appreicated.
>
> Thanks in advance
> Tim Jewell
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
Why not try calling Oracle or thier web site. I'm sure they have some
ideas.
My exerience with Oracle indicates that it's more a function of the size
and complexity of the database that will be a determining factor.
Also, the transaction processing rate will be a factor. The number of
users is not as important as what they are doing.
--
John Gluck (Passport Kernel Design Group)
(613) 765-8392 ESN 395-8392
Unless otherwise stated, any opinions expressed here are strictly my own
and do not reflect any official position of Nortel Networks.
------------------------------
From: Jose Manuel Benitez Sanchez <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.setup,comp.os.linux,comp.os.linux.misc,comp.os.linux.redhat
Subject: [Help:] Advice on SMP machines required
Date: Thu, 08 Jun 2000 17:07:32 +0200
Hi,
We have to buy a new computing server. Of course, it'll be a
machine running Linux. But we haven't reached a consensus regarding
the kind of hardware. It's already decided that we'd like a
multiprocessor SMP machine. The first configuration we had in mind was
a tetra Pentium III. However, our local hardware suppliers say that
the architecture of the Intel Pentium III processor doesn't allow for
a tetra processor configuration. According to them, we'll have to use
Pentium III Xeon processors. Can you provide with some information in
this regard? Do you know of any tetra Pentium III, not Xeon, machine?
Our only problem with the Xeon processors is their high
price. May be I'm wrong, but I think they are overpriced, I mean, I'm
not really sure that the increase in price really pays for the
increase in performance. Have any of you any experiences in this
regard? Do you know of any reliable benchmark for these processors?
In addition, although I've seen some ads offering this kind
of servers (tetra Pentium III Xeon) in some magazines (e.g. Linux
Journal) I'm not aware of any open source software (i.e. compilers,
libraries, etc.) which really takes the best out of the Xeon
processors. Is this appreciation right?
There's also the possibility of using AMD Athlon processors, but
I
think I've read somewhere that the SMP kernel part doesn't work very
well with AMD processors. Can you confirm or reject this point?
Finally, I would thank whatever proposal for parallel
computing configuration (other than beowulf-like clusters). We are
interested in a real single machine, rather than in a group of
independent machines tied together through netword cards.
I'd really appreciate all your feedback.
Thanks a lot in advance!!
Greetings.
Jos� Manuel
--
Jose Manuel Benitez Sanchez e-mail: [EMAIL PROTECTED]
Dpto. de Ciencias de la Computacion e Inteligencia Artificial
E.T.S. Ingenieria Informatica
Universidad de Granada Tel. +34 - 958 - 24 61 43
18071 - GRANADA (Spain) Fax: +34 - 958 - 24 33 17
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.hardware) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Hardware Digest
******************************
