Linux-Hardware Digest #409, Volume #14 Tue, 27 Feb 01 10:13:05 EST
Contents:
Re: Need LOTS of disks: Promise ATA RAID?? (pip)
Notebook HP OB XE3: Problems with network (Accton EN2242) ("Horst Exenberger")
Re: Need LOTS of disks: Promise ATA RAID?? (Jason Clifford)
Overlay with VooDoo 3 (Marc Ariberti)
Voodoo3, OpenGL and Half-Life (Marc Ariberti)
Re: Need LOTS of disks: Promise ATA RAID?? (pip)
Sound - how to configure microphone ("Randy")
Re: Need LOTS of disks: Promise ATA RAID?? (Jonathan Buzzard)
Re: Need LOTS of disks: Promise ATA RAID?? (Jonathan Buzzard)
a7v and R.H 6.2 (Marc de Huu)
Re: DLink DFE-530TX+ NIC OK with Linux ? (Joe Umiker)
Re: USR modem in Red Hat 7.0 (RDetenbeck)
VESA energy saver with a new 2.4.0 kernel? (Otto J. Makela)
Re: Need LOTS of disks: Promise ATA RAID?? (Paul Martin)
Re: HP Deskjet 812C in Linux ("Chris Ripp")
Re: Need LOTS of disks: Promise ATA RAID?? (Paul Martin)
Re: Linux partitioning question (Rod Smith)
Re: Problem with 3com TokenLink Velocity PCI 3c359 (Mike Phillips)
Re: Need LOTS of disks: Promise ATA RAID?? (Joshua Baker-LePain)
----------------------------------------------------------------------------
From: pip <[EMAIL PROTECTED]>
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Tue, 27 Feb 2001 11:11:44 +0000
Jason Clifford wrote:
>
> On 27 Feb 2001, Mark Hahn wrote:
>
> > > If they're going to get heavy use (multiple simultaneous accesses),
> > > then I wouldn't recommend IDE.
> >
> > this is a silly old wives tale. there's nothing "concurrency-challenged"
> > about IDE: all OS's do queue sorting.
>
> You obviously have never tried to implement a busy RAID system using IDE.
> I have and to say the results were unsatisfactory would be akin to saying
> the Atlantic Ocean is a puddle.
>
> For large/busy raid solutions I'd avoid IDE based RAID and opt for a
> hardware based RAID solution using high end SCSI devices.
>
> It's more expensive but unless you don't value your data it's worth it.
How have you found smaller IDE RAID? I was thinking of using a
two IDE disks on different channels and mirroring them and reiserfs'ing
them - would this be a bad idea? Is the performance ok for a small
number
of drives (I was hoping that performance on reads would increase!).
Would prefer the SCSI option of course, but for the cost :-(
I was looking at the new abit KV200-R motherboard (
http://www.motherboards.com/catalog/product_info.php?products_id=266&; )
has anyone tried this? I was concerned about VIA chipset support
(there seem to be constant mentions of this in the kernel mailing list -
I'll need to do some research!).
------------------------------
From: "Horst Exenberger" <[EMAIL PROTECTED]>
Subject: Notebook HP OB XE3: Problems with network (Accton EN2242)
Date: Tue, 27 Feb 2001 12:16:39 +0100
Reply-To: "Horst Exenberger" <[EMAIL PROTECTED]>
Hi,
I installed Linux (SuSE 7.0) on this notebook, but I`ve got problems about
the network. It's seems not to be possible to make it work. Somebody has a
solution?
Thanks
Horst
------------------------------
From: Jason Clifford <[EMAIL PROTECTED]>
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Tue, 27 Feb 2001 12:08:29 +0000
On Tue, 27 Feb 2001, pip wrote:
> How have you found smaller IDE RAID? I was thinking of using a
> two IDE disks on different channels and mirroring them and reiserfs'ing
> them - would this be a bad idea? Is the performance ok for a small
> number
> of drives (I was hoping that performance on reads would increase!).
> Would prefer the SCSI option of course, but for the cost :-(
RAID with independant channel IDE disks is quite good for small
requirements. I have to warn however that I've had issues with software
RAID mirroring recently (as in catastrophic failure).
Last time I looked at it reiserfs had serious compatibility issues with
software raid. That may be resolved now but check before you commit any
data to it.
> I was looking at the new abit KV200-R motherboard (
> http://www.motherboards.com/catalog/product_info.php?products_id=266&; )
> has anyone tried this? I was concerned about VIA chipset support
> (there seem to be constant mentions of this in the kernel mailing list -
> I'll need to do some research!).
My understanding is that there are serious problems with the VIA IDE
chipset support for ATA-100 although not everyone appears to have the
problems. I don't use it here though so I will be selfish and say I don't
care too much.
Jason Clifford
------------------------------
From: [EMAIL PROTECTED] (Marc Ariberti)
Subject: Overlay with VooDoo 3
Date: 27 Feb 2001 12:27:01 GMT
I tried several times to make overlay work (to accelerate VideoLAN)
but it seems that the SDL library can't find overlay with my 3dfx.
Does Voodoo3 support it ?
If yes, how to make it work ?
I use a debian (sid), kernel 2.4.1, XFree 4.0.2, libSDL 1.1.7
--
Marc ARIBERTI
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Marc Ariberti)
Subject: Voodoo3, OpenGL and Half-Life
Date: 27 Feb 2001 12:30:02 GMT
I tried it several times and when I finally make it work, I only had
1fps in OpenGL which is far less than in software mode. Do you
know where is the problem and how to make 3d acceleration work
properly under linux with a Voodoo3
I use one of the latest version of wine, xfree...
--
Marc ARIBERTI
[EMAIL PROTECTED]
------------------------------
From: pip <[EMAIL PROTECTED]>
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Tue, 27 Feb 2001 12:29:57 +0000
Jason Clifford wrote:
> On Tue, 27 Feb 2001, pip wrote:
> > How have you found smaller IDE RAID? I was thinking of using a
> > two IDE disks on different channels and mirroring them and reiserfs'ing
> > them - would this be a bad idea? Is the performance ok for a small
> > number
> > of drives (I was hoping that performance on reads would increase!).
> > Would prefer the SCSI option of course, but for the cost :-(
> RAID with independant channel IDE disks is quite good for small
> requirements. I have to warn however that I've had issues with software
> RAID mirroring recently (as in catastrophic failure).
>
> Last time I looked at it reiserfs had serious compatibility issues with
> software raid. That may be resolved now but check before you commit any
> data to it.
Thanks, yes I'll be cautious as the whole point is for reliability!
Maybe using a motherboard with the built-in IDE hardware RAID would
be a better idea - it is only a bit more expensive anyway.
Thanks
------------------------------
From: "Randy" <[EMAIL PROTECTED]>
Subject: Sound - how to configure microphone
Date: Tue, 27 Feb 2001 07:47:44 +0500
Is there some trick I'm missing to configure my microphone? Everything
regarding sound is working perfectly, but I get nothing out of my
microphone. I also can't find any info ANYWHERE regarding how to get a
microphone to work.
Any ideas?
------------------------------
From: [EMAIL PROTECTED] (Jonathan Buzzard)
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Tue, 27 Feb 2001 08:58:59 +0000
In article <97fefp$d91$[EMAIL PROTECTED]>,
Mark Hahn <[EMAIL PROTECTED]> writes:
>> If they're going to get heavy use (multiple simultaneous accesses),
>> then I wouldn't recommend IDE.
>
> this is a silly old wives tale. there's nothing "concurrency-challenged"
> about IDE: all OS's do queue sorting.
Not true, two IDE drives on the same bus have contention problems. That
said in an IDE RAID system you generally only have one drive on each IDE
bus so it does not matter.
JAB.
--
Jonathan A. Buzzard Email: [EMAIL PROTECTED]
Northumberland, United Kingdom. Tel: +44(0)1661-832195
------------------------------
From: [EMAIL PROTECTED] (Jonathan Buzzard)
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Tue, 27 Feb 2001 09:04:07 +0000
In article <[EMAIL PROTECTED]>,
"C. Newport" <[EMAIL PROTECTED]> writes:
[SNIP]
>
> Be reasonable, in the given context it is marginally possible
> to use IDE, leaving little or no room for expansion and horrid
> performance.
Speeds upto 80MB/s on systems I have personaly made is not horrid
performance. He also might not need expansion, I have had to put
together systems with this sort of storage requirement where the
likely hood of expansion is near zero.
> The website that I cited offers a range of SCSI and Fibre Channel
> arrays, including inexpensive IDE/SCSI hybrids.
>
> In the context of the question, cost is a factor, so I did
> not bother suggesting FC. Using PC hardware there is little benefit
> in using FC rather than SCSI because the throughput is limited by
> the architecture rather than the upper bound of SCSI performance.
> FC would only be usefull in this context if the disk array needed
> to be located too far from the CPU for SCSI cabling.
>
Well this all depends. It might be that with his application a PC based
system is unsuitable. It might be that a FC system hosted off a Unix
box is the only sensible way to go.
JAB.
--
Jonathan A. Buzzard Email: [EMAIL PROTECTED]
Northumberland, United Kingdom. Tel: +44(0)1661-832195
------------------------------
From: Marc de Huu <[EMAIL PROTECTED]>
Subject: a7v and R.H 6.2
Date: Tue, 27 Feb 2001 15:05:52 +0100
Hi,
I have a A7V asus motherboard with 1 cd and 1 cd-writer on the ide-bus
and my hard-disk connected to the ATA-100 thing.
I managed to install R.H 6.2 after some real efforts BUT if I try to
boot without the interactive mode, it dies when starting the cron
daemon. If I boot in interactive mode, choose not to start the cron
daemon, everything works fine. Here comes the killer. If I then start
the cron daemon by hand, the system doesn't crash...
Anyone any idea ??
--
%__________________________________________________________________
%Marc de Huu, Kernfysisch Versneller Instituut, Zernikelaan 25
% 9747 AA Groningen, Nederland. tel: (xx31) 50/3633632
%Auwemalaan 127, 9351 NC Leek, Nederland. tel: (xx31) 594/511546
------------------------------
From: Joe Umiker <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: DLink DFE-530TX+ NIC OK with Linux ?
Date: Tue, 27 Feb 2001 13:53:33 GMT
Sean wrote:
> I currently have a couple of DLink DFE-530TX NICs working OK with my
> SuSE 6.4 setup using the via-rhine driver.
>
> I was going to get some more of these cards but I notice that D-Link
> only seem to have the DFE-530TX+ available now.
>
> Can anyone tell me if this new card works OK with Linux ? If so, what
> driver does it use ? Is it still the via-rhine driver ?
>
> Cheers,
>
> Sean.
Works fine. Use the rtl8139 driver.
------------------------------
From: [EMAIL PROTECTED] (RDetenbeck)
Date: 27 Feb 2001 14:14:06 GMT
Subject: Re: USR modem in Red Hat 7.0
I seem to have solved that problem on my Dell 4100 PC, running Windows ME and
Redhat Linux 7.0. The following is a note that I wrote to myself lest I
forget.
file=/root/notes/modem.txt
rd:2/26/01
Adding a USR/3Com PCI Fax Modem PC5610 to RedHat Linux 7.0
The normal boot/probe process does not properly identify the modem.
(1) Run the program lspci -v to obtain what information the system has about
PCI pnp devices:
{
02:0d.0 Serial controller: US Robotics/3Com 56K FaxModem Model 5610 (rev 01)
(prog-if 02 [16550])
Subsystem: US Robotics/3Com: Unknown device 00d7
Flags: medium devsel, IRQ 3
I/O ports at dff0
Capabilities: [dc] Power Management version 2
}
(2) Set up the serial port /dev/ttyS5 for this device by adding the file
/etc/rc.serial, which will be called by /etc/rc.d/rc.sysinit :
{
# file=/etc/rc.serial
# called by /etc/rc.d/rc.sysinit
# sets USR/3Com 5610 modem parameters for /dev/ttyS5
setserial /dev/ttyS5 uart 16550 port 0xdff0 irq 3 baud_base 115200 spd_vhi
^fourport
setserial -b /dev/ttyS5
}
(3) Link to /dev/modem:
{
ln -s /dev/ttyS5 /dev/modem
}
(4) Test by executing rc.serial (chmod?) and trying to run minicom.
============
I am curious to know whether this simple solution will work for you.
Bob Detenbeck
------------------------------
Crossposted-To:
sfnet.atk.linux,comp.os.linux.setup,comp.os.linux.misc,alt.os.linux.redhat,redhat.config,comp.os.linux.x
Subject: VESA energy saver with a new 2.4.0 kernel?
From: [EMAIL PROTECTED] (Otto J. Makela)
Date: Tue, 27 Feb 2001 14:44:50 GMT
I assume the whole logic behind the VESA energy saving system has
changed, or something, because the settings in XF86Config-4 which
worked well with a older kernel no longer (with a hand-compiled 2.4.0)
switch my monitor to energy saving from the gdm login screen, nor does
the Gnome screen saver ever switch over to energy saving mode though
it has been set to do so. This is a RH7.0 system with XFree86 4.0.1,
the old kernel was the 2.2.16-22 that came with the system.
I got a report of a similar problem from another user, and then again,
someone else told me that it works just fine for him.
How should I proceed to figure out what's wrong?
--
/* * * Otto J. Makela <[EMAIL PROTECTED]> * * * * * * * * * * * * * * * */
/* Phone: +358 40 765 5772, FAX: +358 42 7655772, ICBM: 60N 25E */
/* Mail: Mechelininkatu 26 B 27, FIN-00100 Helsinki, FINLAND */
/* * * Computers Rule 01001111 01001011 * * * * * * * * * * * * */
------------------------------
From: [EMAIL PROTECTED] (Paul Martin)
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: 27 Feb 2001 14:49:21 GMT
In article <97fefp$d91$[EMAIL PROTECTED]>,
Mark Hahn wrote:
>> If they're going to get heavy use (multiple simultaneous accesses),
>> then I wouldn't recommend IDE.
>this is a silly old wives tale. there's nothing "concurrency-challenged"
>about IDE: all OS's do queue sorting.
Some key phrases you might not be aware of: "Scatter gather, otherwise
known as tagged queueing", "Disconnect" and "Post".
A given SCSI drive will usually know better than the OS how to order
its seeks.
--
Paul Martin <[EMAIL PROTECTED]>
------------------------------
From: "Chris Ripp" <[EMAIL PROTECTED]>
Crossposted-To: comp.periphs.printers
Subject: Re: HP Deskjet 812C in Linux
Date: Tue, 27 Feb 2001 08:04:34 -0600
I've got an 812C that I've been using with Linux for some time.
You won't be able to get quality as good as you would with Windows (no way
to fine-tune things for output etc.) but it'll work for just printing docs,
etc.
Use printtool and choose the filter that is also for the 690 series (I
forget which all it includes.)
You'll have to select a couple of the other options (fix stair-step, force
FF, etc.) again I forget which ones exactly (been a while since I set it
up.)
This is all assuming that you can get printtool to detect your lp0....
"Mark Bratcher" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> In article <[EMAIL PROTECTED]>, Jason Noble wrote:
> >I have an HP Deskjet 812C I am trying to setup under Red Hat 6.1 I
> >prints fine under Windows, but Linux will not detect the lp port. I
> >have inserted the line "alias parport_lowlevel parport_pc" to
> >/etc/conf.modules and run "modprobe lp" as root, but printtool still
> >will not detect the lp port. Is there something else I need to do? I
> >had it printing once before, so I know it will work. Any help would be
> >greatly appreciated.
> >
>
> Jason,
> Are you able to get output from your printer by doing the following:
>
> cat /etc/group > /dev/lp0
> echo -e "\f" > /dev/lp0
>
> (Assuming you are on lp0, of course)?
>
> --
> Mark Bratcher
> To reply direct, remove both underscores (_) from my email name
> ---------------------------------------------------------------
> Escape from Microsoft's proprietary tentacles: use Linux!
------------------------------
From: [EMAIL PROTECTED] (Paul Martin)
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: 27 Feb 2001 14:49:21 GMT
In article <97felb$d91$[EMAIL PROTECTED]>,
Mark Hahn wrote:
>> in using FC rather than SCSI because the throughput is limited by
>> the architecture rather than the upper bound of SCSI performance.
>this is rather peculiar, given that it's trivial to knock together
>an IDE system (at $5/GB) that sustains 90 MB/s. while it's certainly
>not difficult to accomplish this with SCSI, it will cost 4x as much.
Is that 90MB/s of linear reading? This sort of access occurs very
rarely on a server in practice (eg. video playout). I'd suggest using
something like multiple simultaneous bonnie++ processes to see how it
performs under a server-like load.
--
Paul Martin <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Rod Smith)
Crossposted-To: comp.os.linux.setup,comp.os.linux.misc
Subject: Re: Linux partitioning question
Date: Tue, 27 Feb 2001 14:54:04 -0000
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
Floyd Davidson <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] (Rod Smith) wrote:
>>Splitting off multiple partitions has several advantages, such as a
>>reduced chance of problems should a runaway process create a too-large
>>file and a reduced chance losing all your data in case of a filesystem
>>problem. IMHO, these are all dwarfed for new users by the near certainty
>>of getting partition sizes wrong, but those who know how big to make
>>their partitions may prefer to split things up.
>
> Since SysV R3 (with symbolic links) it has been almost
> *impossible* to get the partition sizes wrong, and hence there
> is absolutely no need to readjust partitions sizes (which was
> indeed a *serious* admin consideration when installing a SysV R2
> UNIX). Instead entire directories can be placed on other
> partitions and symlinked to any given location.
This is pretty ugly, IMHO, but it does of course work. Part of the
problem is that it cuts into the advantages of having multiple
partitions to begin with. For instance, suppose you normally mount
/usr/local read-only for security reasons, activating read/write access
only when you install a new package or upgrade a new one. When it runs
out of space and the only space left is on /home and link some
directories in /usr/local to somewhere in /home, you lose the read-only
nature of /usr/local for those files.
Furthermore, my objection to lots of filesystem splits is not for
experienced users, but for newbies. Will a newbie know the commands to
use to safely move these files? Probably not. A newbie with incomplete
instructions on doing this could easily produce a non-bootable system
and have to reinstall (or do SERIOUS maintenance, which this newbie
would also be unqualified to do).
--
Rod Smith, [EMAIL PROTECTED]
http://www.rodsbooks.com
Author of books on Linux & multi-OS configuration
------------------------------
From: Mike Phillips <[EMAIL PROTECTED]>
Subject: Re: Problem with 3com TokenLink Velocity PCI 3c359
Date: Tue, 27 Feb 2001 14:57:21 GMT
No it won't, the 3c359 uses a different chipset from the 3c339 adapter
(which will work with the tms380 driver). A beta driver for the 3c359 is
available from http://www.linuxtr.net. This will be submitted for
inclusion in the kernel soon.
Mike Phillips
Linux Token Ring Project.
http://www.linuxtr.net
On Tue, 27 Feb 2001, Jan Kov=E1=F8 wrote:
> I hate token ring. Had the same problems. I have bought SysKonnect PCI
> network adapter for linux and left the 3Com card for the PCs with Windows=
9x.
> But teoreticaly it may work. Recomplile the kernel and use tms380 module.=
It
> should support Compaq 4/16 TR PCI, SysKonnect TR4/16 PCI, Thomas-Conrad
> TC4048 PCI 4/16, 3Com Token Link Velocity.
>=20
> Good luck. Post a note if you have been successful.
>=20
> TNX
>=20
>=20
>=20
> Moritz Petersen <[EMAIL PROTECTED]> p=ED=B9e v diskusn=EDm
> p=F8=EDsp=ECvku:90lc4s$l8$[EMAIL PROTECTED]
> > Hello,
> >
> > does anyone have experiences with a 3com TokenLink Velocity PCI with 3c=
359
> > chipset?
> > With my SuSE 7.0 Professional distribution there were no default driver=
s
> for
> > this card and unfortunateley I cannot find any '3c359.o' file in the
> > internet anywhere... neither at the SuSE support database nor at the 3c=
om
> > support pages.
> >
> > Any help will be appreciated, thanks
> >
> > Moritz Petersen
> >
> >
>=20
>=20
>=20
------------------------------
From: Joshua Baker-LePain <[EMAIL PROTECTED]>
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: 27 Feb 2001 15:02:17 GMT
In comp.os.linux.hardware Jonathan Buzzard <[EMAIL PROTECTED]> wrote:
> In article <97e893$6no$[EMAIL PROTECTED]>,
> Joshua Baker-LePain <[EMAIL PROTECTED]> writes:
>>
>> I'm in the market for disk space of this magnitude at the moment as
>> well. While the above statement used to be the case, I have heard
>> a number of *very* informed opinions in support of IDE-SCSI RAID
>> towers, such as the ones at www.zero-d.com. It's a tower/rack-mount
>> full of IDE disks that'll do RAID 0, 1, or 5 (possibly 10, I don't
>> recall at the moment). They have hot swappable drives, power supplies
>> and fans. Finally, they connect to the host via a SCSI (typically UW or
>> U2W) connection, so are OS independent.
> These are good but the write perfomance on RAID 5 is a bit sucky and
> the read is not too hot either. You really need to do your homework
> first.
Ahem. I've been doing lots of homework on this. The original poster
never stated blazing performance as a requirement. I am quite aware
that IDE is not going to read/write as fast as SCSI. Speed is not
always a requirement. In my case, I need a pretty decent sized array
(.5TB), but all users will be accessing it via NFS on a 100 Mbit network.
The network is far more likely to be the bottleneck than is the array.
And I don't have ~$15K to spend on a SCSI RAID of that size, nor do I
need to.
> Yep, but the IDE arrays don't go as fast as the SCSI ones and the
> zero-d website says as much.
Again, speed isn't always a requirement. For big arrays where speed
isn't critical, IDE/IDE-SCSI RAID are very viable options.
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
------------------------------
** 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 by posting to comp.os.linux.hardware.
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
******************************
