Linux-Hardware Digest #413, Volume #14 Wed, 28 Feb 01 04:13:06 EST
Contents:
Re: Should I abandon SCSI? (J.B. Nicholson-Owens)
Re: Need LOTS of disks: Promise ATA RAID?? ("Leo")
Re: Best RAID controller for Linux ("Leo")
Re: USR CP5610A Modem in RH 7.0 ("Crazy")
lm_sensors: Only got voltage. Where's CPU temp? (Dan Smith)
Re: Initio's lack of Windows 2000 support (Peter Seebach)
Re: Initio's lack of Windows 2000 support ([EMAIL PROTECTED])
Re: Boot Redirection ([EMAIL PROTECTED])
New SCSI drive partition information different for fdisk and cfdisk???? ("Pete
Willemsen")
Re: VIA IDE problems (was Re: Need LOTS of disks: Promise ATA RAID??) (Kenneth
R�rvik)
Promise FastTrak100 - OK, now what? (iQXth)
Netgear EA201? ("Jan Israelsson")
Re: GLX driver for Riva TNT in RedHat (Marcus =?ISO-8859-1?Q?H=E4gglund?=)
plug modem at paralell device? (Thilo Riessner)
video card problem ("albert")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (J.B. Nicholson-Owens)
Crossposted-To: comp.periphs.scsi
Subject: Re: Should I abandon SCSI?
Date: 28 Feb 2001 05:16:01 GMT
Donovan Rebbechi wrote:
> RAID 0 is great for high performance desktops, but if you want a
> storage solution that's also *reliable*, it's useless.
I wouldn't say it's useless--that depends on the data stored on the volume
and how well the site has prepared for failure. But I agree--RAID 0 is a
high price to pay for failure. A switch to RAID 3 or 5 would help a single
drive failure (if I recall correctly, RAID levels 3 & 5 use one drve in a
set for parity data and can restore any single drive failure the instant the
drive is replaced and low-level formatted, if necessary. If the parity
drive fails it can be rebuilt from the other drives in the set when the
parity drive is replaced).
> (BTW, shouldn't RAID 0 be "AID 0" ? THe R doesn't really apply)
Heh, good point.
------------------------------
From: "Leo" <[EMAIL PROTECTED]>
Crossposted-To: uk.comp.os.linux
Subject: Re: Need LOTS of disks: Promise ATA RAID??
Date: Wed, 28 Feb 2001 05:24:21 GMT
Holy huge thread!
Sorry, I couldn't refuse to apply my .02
Guys guys guys... Each technology has it's place. IDE is definitely your
low cost point of entry controller. I'd say good for some people looking
for inexpensive, and perhaps lower performance/scalability requirements.
SCSI controllers, and SCSI RAID controllers definitely enable connection of
larger numbers of disks and offload I/O specific tasks from the CPU to the
controller. Yes, there are some companies out there making some low cost
IDE array type devices. Perhaps there are fits for this, but call me
skeptical about trusting my business critical servers to an IDE RAID (or IDE
RAID converted back to SCSI! seems like there would have to be some
performance hit with that translation?). BTW - Are any of these IDE RAID
solutions endorsed by the Raid Advisory Board? I haven't looked.
Someone earlier mentioned fibre channel as an alternative. If someone is
worried about cost enough to consider launching some critical app with an
IDE RAID solution of any magnitude to save $, you can forget FC. The HBA's
alone will probably cost more than the entire IDE solution :) Oh yeah, and
on the fibre note-- someone mentioned "Not unless distance is an issue".
That's very smart and true. Fibre channel DOES NOT always outperform SCSI.
Remember, SCSI uses many parallel paths where fibre is a serial type
delivery that passes SCSI protocols. So, although it's 'faster', due to the
encapsulation and de-capsulation if you will, of the SCSI protocol to pass
it over fibre channel depending on your I/O block size it may not be faster.
Typically small block I/O is faster with SCSI. Typically when you get into
block sizes of 16 or 32K, fibre will win. Point is, it's more expensive and
not always faster-- but can be. Also, fibre rules for distance and
connectivity... you can fan out through a switch fabric and connect all
kinds of servers.
Anyway, for the guy who originally asked about his IDE situation-- (you sure
opened a can of worms! :)) IDE could be fine. It depends on the
requirements including availability, performance, connectivity, scalability,
and ease of management.
I don't think IDE is for anything high end that needs the requirements
stated above. I don't see EMC running out to put IDE drives in their
enterprise storage systems, or Hitachi, or IBM, or Compaq, (MTI, Sun,
NetApp, blah blah blah....) it's just not the right tool...
Hey, I love IDE. Using it in the machine I'm typing on right now. But
guess what, if it died (you would be spared my email!) it wouldn't be
"critical". Our farm of web servers however serving our own and our hosting
client's sites is a different story... and worth the dough for ICP array
controllers and speedy fast SCSI drives.
NAS? NAS is cool too, provided you don't need performance of channel speed.
NAS rules for serving content without replicating it for multiple
servers/environments. Cool for file sharing... making an NT share (CIFS) to
the same data that is being served to UNIX via NFS. Or making the data
available over a WAN. I wouldn't use it for a busy critical database
environment though.
EVERYTHING HAS IT'S PLACE. Understand your requirements and deploy the
right technology. Use the right tool. Using IDE for mission critical
storage requirments is like bringing a knife to a gun fight. Using
Ultra160MB/s with spanky-fast 10K drives with all kinds of redundacy for
something non-critical reminds be of that credit card commercial where they
are trying to get the cat out of the tree "It's a Calico, we need more air
support!". What about a ladder? yeah, IDE's your fit.
Leo J. Squire
www.icp-order.com
High Availability, Now Highly Available
"Steve Wolfe" <[EMAIL PROTECTED]> wrote in message news:otah79.o66.ln@helix...
> > this is a silly old wives tale. there's nothing
"concurrency-challenged"
> > about IDE: all OS's do queue sorting.
>
> I won't get into arguing about symantics, but I've used IDE and SCSI in
> areas where there are multiple programs accessing the disk at once. SCSI
> blows away IDE in terms of responsiveness and throughput. That's all I
need
> to know.
>
> steve
>
>
>
------------------------------
From: "Leo" <[EMAIL PROTECTED]>
Crossposted-To: linux.dev.raid
Subject: Re: Best RAID controller for Linux
Date: Wed, 28 Feb 2001 05:41:14 GMT
That's great feedback...
I can only say as a consultant who has tried a lot of controllers in
multiple OS's (Linux, and Novell/NT) ICP has been the best. I'm certainly
not slamming others, I've successfully deployed many other array controllers
including Adaptec, Compaq SMART Array Controllers, HP NetRAID, etc. But,
ICP has been the best overall. Best performing, best support, and best ease
of setup our consulting team has experienced.
Since we started www.icp-order.com, I've asked customers where they heard of
ICP and why they made an ICP decision. One of the latest responses I
received was "We saw ICP at Linuxworld and were impressed. After having
numerous issues with Adaptec (and he mentioned driver related issues) we
decided to try ICP.
Actually, ICP is the ONLY Tier 1 RAID controller approved by Redhat. Check
out http://www.redhat.com/support/hardware/intel/61/rh6.1-hcl-i.ld-5.html
and scroll to section 5.6.
Sincerely,
Leo J. Squire
www.icp-order.com
High Availability, Now Highly Available
"Vincent Fox" <[EMAIL PROTECTED]> wrote in message
news:97e0ko$s6b$[EMAIL PROTECTED]...
> In <97bsr5$40k$[EMAIL PROTECTED]> Juergen Sauer <[EMAIL PROTECTED]>
writes:
>
> *snip*
>
> I've started using the 3Ware IDE RAID controller and been
> very impressed with it so far. Using a pair of IBM 75GXP drives
> in mirroring mode, disk reads are very fast. The mirroring
> works as advertised. I can "fail" a drive by unplugging it
> and it keeps running. Rebuilding onto the failed drive
> to return to full mirroring functionality can be done while
> the system is running. Sweet.
>
> Only weird thing I note is that if I put my Tekram D390U3W card
> in with a Seagate X15 drive it seems I cannot access it. I think
> the 3Ware and the Tekram card conflict.
>
>
> --
> "Who needs horror movies when we have Microsoft"?
> -- Christine Comaford, PC Week, 27/9/95
------------------------------
From: "Crazy" <[EMAIL PROTECTED]>
Subject: Re: USR CP5610A Modem in RH 7.0
Date: Tue, 27 Feb 2001 22:58:53 -0700
I have this same modem and am using Redhat 7.0 with kernel 2.2.16. Here is
what I had to do to kind of get it working, but bear in mind that I am an
EXTREMELY new user to Linux and am still trying to figure this out:
login as root
issue the command 'lspci -v'
somewhere on the resulting list, you will find your modem listed, what you
need to get from that listing is the port and irq numbers.
issue the command 'setserial /dev/ttyS1 port0x1430 irq 10 autoconfig', and
insert your port and irq numbers in the appropriate spots. This is also
where I am confused about how to figure out which serial port is free so you
can assign your modem to it, I just happened to use /dev/ttyS1 cause I don't
have a clue what else to use, and it happened to work, so if you know how to
find a free serial port, make sure you use the appropriate serial port
number instead of /dev/ttyS1.
If you get this far without getting an error message, then most likely your
modem is configured okay. If you get an error message, then again I really
don't know what do to cause I'm clueless hehe.
If you don't have any error messages, then issue the command 'ln /dev/ttyS1
/dev/modem' again using the appropriate serial port if its different than
/dev/ttyS1. This command creates a symbolic link of the serial port to
'modem', making it easier to configure in your dialing program.
Once this is all done, you need to test it out in your dialer program (I use
rnPPP dialer or something like that), and if it works, then you are supposed
to add the above 'setserial' command and 'ln' command to a file called
'rc.local' (you just add them to the bottom of the file and save it), but I
tried this and I still have to reconfigure my modem each time I bootup, so
I'm doing something wrong there.
Also, someone told me that this is all fixed in kernel 2.4, and that all I
have to do is download the new kernel and compile it and its all fixed, but
this is where I am really really clueless, so if anyone can help me out with
that, I would be EXTREMELY grateful.
Thanks,
Erik
"Nader" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
My USR modem is working under 2.2.16. In fact, it worked under 2.2.10 until
I encountered a PCI IRQ sharing problem.
Have you tried your modem in 2.2.16? What doesn't work?
Edward Stammer wrote:
> I have RH 7.0 installed, but to my dismay, the modem I ordered for using
> with Linux supposedly only works in Kernel 2.3 or higher. I would love to
> update my RH 7.0, kernel 2.2.16, to kernel 2.4, but without a modem, it is
> kind of hard.
>
> If anyone has this modem working in Linux kernel 2.2.16 or less, please,
let
> me know how you did it.
>
> Ed
------------------------------
From: Dan Smith <[EMAIL PROTECTED]>
Subject: lm_sensors: Only got voltage. Where's CPU temp?
Date: 28 Feb 2001 00:40:19 -0500
I am using the lm_sensors package to monitor my board. My board has an LM78 and LM75
chip on it. As I understand, the LM78 is for voltage stuff mostly, and the LM75 is
for temps. Well, even if I insert the lm75 module, I don't see any other info.
Can someone help??
--Dan
------------------------------
Crossposted-To:
comp.unix.bsd.freebsd.misc,comp.unix.bsd.netbsd.misc,comp.os.linux.misc,comp.periphs.scsi,comp.sys.ibm.pc.hardware.misc,comp.sys.ibm.pc.hardware.storage,microsoft.public.win2000.hardware
Subject: Re: Initio's lack of Windows 2000 support
From: [EMAIL PROTECTED] (Peter Seebach)
Date: 28 Feb 2001 06:11:42 GMT
In article <[EMAIL PROTECTED]>, Tim Buck <[EMAIL PROTECTED]> wrote:
>I just found out the hard way that you do not support your INI-9100UW
>SCSI adapter under Windows 2000. I am dumbfounded by your decision to
>not support this product under Windows 2000.
Why? Huge development cost, and *ZERO* sales. They aren't selling this
card anymore; why should they write drivers for it now?
Anyway, if you want to blame someone, blame the folks who decided that they
needed to invent a brand new, incompatible, driver model for their "new"
system.
-s
--
Copyright 2001, all wrongs reversed. Peter Seebach / [EMAIL PROTECTED]
C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon!
Consulting & Computers: http://www.plethora.net/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Initio's lack of Windows 2000 support
Crossposted-To:
comp.unix.bsd.freebsd.misc,comp.unix.bsd.netbsd.misc,comp.os.linux.misc,comp.periphs.scsi,comp.sys.ibm.pc.hardware.misc,comp.sys.ibm.pc.hardware.storage,microsoft.public.win2000.hardware
Date: Wed, 28 Feb 2001 06:32:16 GMT
And no more posting into the wrong newsgroups!
In comp.unix.bsd.netbsd.misc Dave Uhring <[EMAIL PROTECTED]> wrote:
> And it will cost you and your company less to discard your incompatible OS
> and switch to Linux, BSD or Solaris. No more seat licenses, no more
> scheduled "restarts", no more blue screens of death.
> Dave
> Tim Buck wrote:
>> I sent this to Initio a few minutes ago. I thought some of you might be
>> interested in it.
>>
>> ===== begin rant =====
>> I just found out the hard way that you do not support your INI-9100UW
>> SCSI adapter under Windows 2000. I am dumbfounded by your decision to
>> not support this product under Windows 2000.
>>
>> For a company who appears to pride themselves on their cross-platform
>> support, who provides drivers for their hardware for (most) Windows
>> platforms, Linux, BSD, MacOS, UnixWare, and Solaris, the reasoning
>> behind this decision to abandon the 9100UW is unfathomable.
>>
>> I understand and agree that you should not be producing the 9100UW
>> anymore -- Ultra 2 SCSI, LVD, and all the other latest buzzwords make
>> Ultra Wide SCSI obsolete now. I have no problem with that; I do have a
>> problem with you forcing your customers to discard perfectly good,
>> functioning hardware, simply because they wish to upgrade their
>> operating system. You're the only company I've seen do this; for
>> example, I have yet to find a network card for which there's no Windows
>> 2000 driver.
>>
>> What's especially galling about your decision is that you consider the
>> 9100UW an "end of life" product which you no longer support, yet it's
>> still for sale in your online store, as of today!
>>
>> Our company first began buying Initio SCSI adapters as a low-cost,
>> high-performance alternative to expensive Adaptec adapters. We will no
>> longer do so because of your ridiculous decision. It would cost you very
>> little in development time to port your existing Windows NT 4.0 driver
>> over to Windows 2000; it will cost you a lot in lost future sales
>> because you haven't done so.
>>
>> Timothy Buck
>> [EMAIL PROTECTED]
--
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Boot Redirection
Date: Wed, 28 Feb 2001 01:55:59 -0800
Juergen Pfann wrote:
>
> [EMAIL PROTECTED] wrote:
> >
> > I've written a custom boot sector to a floppy disk which when booted, will
> > load the MBR from hdc1. I have MS-DOS on hda1 and Linux on hdc1. The purpose
> > of the disk is to allow me to choose between MS-DOS and Linux when booting.
> > I've tried modifying lilo.conf, but it didn't work.
> >
> > (...)
> > Is there a way for a boot sector on fd0 to make the BIOS or MBR of hdc1 load
> > the system from hdc1?
>
> It's much much simpler than that. For a boot floppy without LILO, just
> to
> boot a linux system from hard disk, you only need to transfer that
> system's
> kernel to the floppy disk :
> 'dd if=/boot/vmlinuz of=/dev/fd0'
> Voila...
> This method assumes that the kernel file we're talking of is configured
> to use the root fs of the system it is compiled with (/dev/hdc1 in our
> case)
> - which is normally the case with a self-compiled kernel, but not
> necessarily if you still use your distro's original installed kernel.
> To assure this, you can issue the command "rdev /boot/vmlinuz /dev/hdc1"
> before the above-mentioned transfer, or "rdev /dev/fd0 /dev/hdc1"
> afterwards, or both if you like.
> BTW : I hope you read at least the relevant man pages (here : "dd"
> and "rdev") - especially "dd" might be a little bit dangerous in the
> hand of someone not 100% knowing what he's doing...
> This method has at least one disadvantage : There's no prompt to enter
> options/arguments to the kernel - just loading the kernel into memory,
> finding the root fs and mounting it, and so on.
> But it's nice in situations where you know your linux fs is o.k, but
> your MBR is not...
>
> Juergen
Is there another way? Due to the size of the kernel, it would add about
a couple of minutes to the boot process because it would have to read
nearly all of the diskette during booting.
------------------------------
From: "Pete Willemsen" <[EMAIL PROTECTED]>
Subject: New SCSI drive partition information different for fdisk and cfdisk????
Date: Wed, 28 Feb 2001 00:05:14 -0700
Hello.
I recently installed Linux (Debian woody) on a P4 machine that has two SCSI
disks. Linux installed well on the second disk (/dev/sdb) which is an IBM
Ultrastar 18GB SCSI disk. Everything seemed perfectly fine with the system
until I attempted to look at the partitions with PartitionMagic.
PartitionMagic reported that the Linux partition had an error. So, I booted
Linux, ran fdisk and received the following report from "fdisk -l /dev/sdb":
----
Disk /dev/sdb: 1 heads, 35565080 sectors, 1 cylinders
Units = cylinders of 35565080 * 512 bytes
Device Boot Start End Blocks Id System
/dev/sdb1 * 1 1 17652064+ 83 Linux
Partition 1 has different physical/logical beginnings (non-Linux?):
phys=(0, 1, 1) logical=(0, 0, 64)
Partition 1 has different physical/logical endings:
phys=(1023, 255, 63) logical=(0, 0, 35304192)
Partition 1 does not end on cylinder boundary:
phys=(1023, 255, 63) should be (1023, 0, 35565080)
----
I never noticed this before, so I decided to run cfdisk. It showed the
partition as being correct:
---
sdb1 Boot Primary Linux ext2
18075.75
---
and the cfdisk table print produced:
---
Partition Table for /dev/sdb
---Starting--- ----Ending---- Start Number of
# Flags Head Sect Cyl ID Head Sect Cyl Sector Sectors
-- ----- ---- ---- ---- ---- ---- ---- ---- -------- ---------
1 0x80 1 1 0 0x83 255 63 1023 63 35304129
2 0x00 0 0 0 0x00 0 0 0 0 0
3 0x00 0 0 0 0x00 0 0 0 0 0
4 0x00 0 0 0 0x00 0 0 0 0 0
---
My big question is why are fdisk and cfdisk reporting different information.
Which is correct? My Linux installation is working, but this anomaly in the
partition table bothers me. Might anyone have any information on why this
is occuring?
Many thanks for your time.
Pete Willemsen
University of Utah
------------------------------
Crossposted-To: uk.comp.os.linux
Subject: Re: VIA IDE problems (was Re: Need LOTS of disks: Promise ATA RAID??)
From: [EMAIL PROTECTED] (Kenneth R�rvik)
Date: Wed, 28 Feb 2001 08:31:34 GMT
[EMAIL PROTECTED] (Tony Houghton) wrote in
<[EMAIL PROTECTED]>:
>> 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.
Runs fine with me with udma5. 2.4.2 is supposed to have a fix for the
vt82c686 driver - according to the changelog. This may solve the problem?
--
Kenneth R�rvik 91841353/22950312
Nordbergv. 60 A [EMAIL PROTECTED]
0875 OSLO home.no.net/stasis
------------------------------
From: iQXth <[EMAIL PROTECTED]>
Subject: Promise FastTrak100 - OK, now what?
Date: Wed, 28 Feb 2001 08:38:03 GMT
I installed Red Hat Linux 7.0 using the Promise drivers disk to
install to a RAID 0 striped array.
The problem is when it configures itself, it doesn't use the 'lba32'
option in '/etc/lilo.conf'. My question is, how do I get it there and
get lilo to apply the changes to the MBR of the array?
The boot disk that was created during install is useless. It gives an
error of '0x10'. I can boot the CD with the 'linux rescue' option, but
then '/dev/sda' is not there. I can also boot the CD with 'linux
rescue dd' option. The device '/dev/sda' is still not there.
Can someone please help me with this? I somehow need to get the driver
loaded to recognize the Promise controller and then mount '/dev/sda3'
as the root partition. If I can get that, I'm sure I can take it from
there.
--- iQXth ---
Please respond to this thread
or post with 'ulvfboqj' in the
subject to get my attention.
------------------------------
From: "Jan Israelsson" <[EMAIL PROTECTED]>
Subject: Netgear EA201?
Date: Wed, 28 Feb 2001 09:21:49 +0100
Hi all,
Can someone please explain how to get the Netgear EA201 ethernet card (ISA)
to work on my IBM PC running RedHat 7.0? As I'm new to Linux please take it
step by step.
Thanks, Jan.
------------------------------
From: Marcus =?ISO-8859-1?Q?H=E4gglund?= <[EMAIL PROTECTED]>
Subject: Re: GLX driver for Riva TNT in RedHat
Date: Wed, 28 Feb 2001 09:55:08 +0100
I've also got an ASUS V3400... 0.9.6 seems to work for some programs,
like some of the demos delivered with Mesa. It doesn't support
dynamic-texturing or something like that (can't really remember what
it's called).
I finally got it to support AGP but although the driver /proc/nv/...
says that sideband is supported it doesn't use it?
Have you tried to use 0.9.6?
Harri Haataja wrote:
> Marcus H�gglund wrote:
>
>> I "uninstalled" (rpm -e) the preinstalled Mesa and disintegrated every
>> GL lib on the system, then reinstalled the Nvidia package and now it
>> seems to work. :-)
>
>
> It works? On a regular TNT_1_ card?
> What card is that?
>
> (I had an ASUS V3400 and 0.9.5 failed in many mysterious ways)
------------------------------
From: Thilo Riessner <[EMAIL PROTECTED]>
Subject: plug modem at paralell device?
Date: Wed, 28 Feb 2001 09:49:00 +0100
Hi,
in my embedded system I need a PC104 analog modem. But the only one I can get
emulates itself on the paralell device. Is it possible to uses mgetty (or
mingetty) with a AT compatible modem on /dev/lp0?
Many thanks in advance for any help.
Thilo
------------------------------
From: "albert" <[EMAIL PROTECTED]>
Subject: video card problem
Date: Wed, 28 Feb 2001 17:03:02 +0800
Trident 975 agp video card cann't installed on redhat.
------------------------------
** 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
******************************
