Linux-Hardware Digest #6, Volume #14 Sat, 9 Dec 00 14:13:08 EST
Contents:
System freeze ([EMAIL PROTECTED])
Re: Linux + big HDD partition ("Gene Heskett")
Re: ARGGGHHH!! Slack7.1 ATI AllinWonder Pro Quake2 and glx ("Jason Harley")
Re: Linux + big HDD partition ("Lee")
Running a headless box ("Mark Riehl")
Re: Format a linux hard drive for Win98 ([EMAIL PROTECTED])
Re: What is the command to . . . ? (Allen Wong)
Re: Running a headless box (Michael Burian)
Re: large file problem on burned cd (Lindenthal)
Re: ECC RAM supported by Linux? (jens)
Re: Trying to port a hacked CMI8330 audio setup from RH6.2 to Potato... (walt)
fasttrak66 ("Enrico Ng")
harddrive upgrade ("Lenny Shovsky")
Re: please help !!!! Format hd problem (David)
Re: Running a headless box (ds1436)
Re: harddrive upgrade (Ned Latham)
which ram? ("eftech_1-=-")
Re: ZIP Plus - need imm.tar.gz (Dances With Crows)
Re: ECC RAM supported by Linux? (Carlos)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: System freeze
Date: Sat, 09 Dec 2000 12:57:34 GMT
I have a box that I'm setting up remotely which has twice died (the last few
days). In both instances all connections drop and running nmap shows that
port 80 is still open - but cannot connect to it. I asked for the machine to
be rebooted the first time and it worked for a few days until now.
Because I am doing this remotely, I cannot see what (if any) error messages
are on screen or in logs. I can think of two possibilities why this crash
happens.
(1). The first time it crashed I was setting up hdparm with -d1 -X34
parameters and it died within seconds - no prompt returning but the telnet
session would still echo characters sent. Since it was rebooted, I had set
only -d1 in hdparm and it worked... but if this drive/chipset doesn't like
dma can it crash after 24hrs or so - not immediately?
(2). The other thing I can think of is a memory leak segfaulting something.
It still sounds odd that all services are dropped like this and no ports open
except port 80.
Anyone have any ideas?
It's a P500 with 64Mb RAM hda: FUJITSU MPE3102AH B3, 9641MB w/2048kB Cache,
CHS=1229/255/63 I think it has an MVP3 or 4 chipset (I think) and after doing
a bit of research it seems this chipset has problems with DMA. It might not
have this chipset but others who do and run hdparm -d1 -X34 get a freeze
instantly too (or at least the few posts I saw).
It has Slackware 7.0 with the following updates:
sysklogd 1.4
apache 1.3.14
BIND 8.2.2-P7
perl 5.6.0
glibcso + glibc 2.1.3
descrypt
I installed these without 'init 1' (obviously I cannot remotely) but rebooted
- I imagine this should still be fine...?
It also has lastest OpenSSL, ssh2, qmail + related packages
I doubt it matters but it has a seperate partitions for:
swap
/
/usr
/home (nosuid)
/tmp (nosuid)
/var
Hopefully I can find out what the problem is instead of rebooting it
constantly.
I am almost convinced it is the dma problem because it never froze before I
set it - but I would have thought if that was the problem then it would crash
as soon as it was set.
The box still has 2.2.13 kernel that has not been recompiled, which may not
have CONFIG_IDEDMA_AUTO enabled and causes DMA to not work...? I was planning
to recompile it after I have everything else working perfectly.
Thanks in advance.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
Date: 9 Dec 2000 8:8:49 -0500
From: "Gene Heskett" <[EMAIL PROTECTED]>
Subject: Re: Linux + big HDD partition
Unrot13 this;
Reply to: <[EMAIL PROTECTED]>
Gene Heskett sends Greetings to Mike ;
> TROUBLE: Linux 2.2.16. I have IDE-HDD 13G. From 600 to 1600
> cylinder - not root Linux partition (83). If data on this partition
> is between 600 and 1024 cylinder - all right. But if i try write
> data to cylinder > 1024 - "No free Space Left". FDISK utility show
> all properly - all 1600 cylinders, but DF utility shows only 3G
> from 8G partition. Q: What I should do to solve this problem?
ISTR there were some big disk patches backported to 2.2.16. Snoop
around on kernel.org, or maybe on rufus. I have a 13g here, but thats
never given me any problems that I know of. No partitions over 6 gigs
though.
Cheers, Gene
--
Gene Heskett, CET, UHK |Amiga A2k Zeus040, Linux @ 400mhz
email gene underscore heskett at iolinc dot net
#Amiga based X10 home automation program EZHome, see at:#
# <http://www.thirdwave.net/~jimlucia/amigahomeauto> #
ISP's please take note: My spam control policy is explicit!
#Any Class C address# involved in spamming me is added to my killfile
never to be seen again. Message will be summarily deleted without dl.
This messages reply content, but not any previously quoted material, is
� 2000 by Gene Heskett, all rights reserved.
--
------------------------------
From: "Jason Harley" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.slackware
Subject: Re: ARGGGHHH!! Slack7.1 ATI AllinWonder Pro Quake2 and glx
Date: Sat, 9 Dec 2000 10:18:51 -0400
Howdy;
I'd suggest an upgrade to X4.0.1, removing the glide_v3 (seeing you don't
have a voodoo 3 with glide support), and the UTAH-glx module (it's part of
X4.x.x anyway).
After the new X is installed, make sure you are using the proper x-server
for your card (r128, I believe), or if you're running a framebuffer try
adding the r128 information to the accel section of your config file.
I hope this helps a little bit,
"dave" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> How the heck can I get all these working together so I can play quake2
> in gl mode?
> I installed the latest version of mesa and glide_v3 and I tried the
> utah-glx driver too.
> I spent about 8 hours on this last night and accomplished nothing. Does
> someone have
> a similar setup and want to relay what all they did to get it up and
> running.
>
> I have the following:
> pentiumII 350
> 128 MB ram
> Ati All-in-Wonder Pro AGP
> quake2 demo
>
>
> Quake2 works fine in software mode, but who the hell wants that?
>
>
> Thanks in advance
>
> Dave
>
------------------------------
From: "Lee" <[EMAIL PROTECTED]>
Subject: Re: Linux + big HDD partition
Date: Sat, 09 Dec 2000 14:21:55 GMT
In article <[EMAIL PROTECTED]>, "Mike" <[EMAIL PROTECTED]>
wrote:
> TROUBLE: Linux 2.2.16. I have IDE-HDD 13G. From 600 to 1600 cylinder -
> not root Linux partition (83). If data on this partition is between 600
> and 1024 cylinder - all right. But if i try write data to cylinder >
> 1024 - "No free Space Left". FDISK utility show all properly - all 1600
> cylinders, but DF utility shows only 3G from 8G partition. Q: What I
> should do to solve this problem?
>
>
>
>
Hello, I had a similar problem. I have a 20G Maxtor hd. fdisk could only see a
small part of the drive. I used cfdisk and it was able to see all the space. I
also partitioned the disk in to 3 separate partitions. I used a 128MB swap
partition, I gave 8GB to my root file system, and I gave the rest to my
home directory.
------------------------------
From: "Mark Riehl" <[EMAIL PROTECTED]>
Subject: Running a headless box
Date: Sat, 09 Dec 2000 14:28:39 GMT
All,
I'd like to run a headless box (no monitor or keyboard) sitting on the floor
for backups. It will be an older 486 with a LAN card and a video card.
Is there a trick that will let me boot w/o a keyboard? Or, do I need to
boot with the keyboard and then pull it off?
Thanks,
Mark
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Format a linux hard drive for Win98
Date: Sat, 09 Dec 2000 15:02:08 GMT
After removing them with linux fdisk and before removing them with dos
fdisk, try doing an fdisk /mbr. This should write a new mbr.
In article <[EMAIL PROTECTED]>,
Pankaj Chauhan <[EMAIL PROTECTED]> wrote:
> >I need to format a hard drive that has linux on it to re-use it for
win98.
> >When I use a win98 boot disk the format command won't recognise the
drive
> >structure. The format command fails.
>
> Hi,
>
> What about using fdisk to remove all the partitions that you have?
> First use linux fdisk if that's possible, o/w use dos fdisk to wipe
> out any partitions that it may see. Then try to format after that.
>
> There are two kinds of format that windows format program offers I
> think, one is quick format and another if full format. Do both of
> them fail?
>
> Another idea might be to do dd if=/dev/zero of=/dev/hdX (X is a, b, c
> or something depending upon which device it is linux or if it is SCSI
> then /dev/sdX??), which will write zeroes all over the disk! Then use
> format program.
>
> --
> - Pankaj
>
> ----------------------------------------------------------------------
> "Great ideas are like internet domain names, the ones
> I get are already taken up by Knuth, Neumann, Turing
> and the likes....."
> ----------------------------------------------------------------------
>
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Allen Wong)
Subject: Re: What is the command to . . . ?
Crossposted-To:
alt.os.linux.mandrake,alt.os.linux.slackware,alt.uu.comp.os.linux.questions,comp.os.linux.development.system,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.setup
Date: Sat, 09 Dec 2000 15:22:07 GMT
In alt.os.linux.slackware Markus Amersdorfer <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED] wrote:
>>
>> What is the command to find the occurance of a string in a group of
>> files in a directory?
> find . -name '*.txt' -exec grep "Hello World" {} \;
This works, but it's alot slower than "find . -type f -name '*.txt' -print |
xargs grep "Hello World".
Allen
--
Linux: If you're not careful, you might actually learn something.
7:00am up 1 day, 10:59, 9 users, load average: 3.09, 3.04, 3.00
------------------------------
From: Michael Burian <[EMAIL PROTECTED]>
Subject: Re: Running a headless box
Date: Sat, 09 Dec 2000 16:32:41 +0100
Mark Riehl wrote:
>
> All,
>
> I'd like to run a headless box (no monitor or keyboard) sitting on the floor
> for backups. It will be an older 486 with a LAN card and a video card.
>
> Is there a trick that will let me boot w/o a keyboard? Or, do I need to
> boot with the keyboard and then pull it off?
>
> Thanks,
> Mark
Some BIOS have got an Option "HALT ON", set it to "NO ERROR"
or something like that and see if it works.
------------------------------
From: Lindenthal <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help
Subject: Re: large file problem on burned cd
Date: Sat, 09 Dec 2000 16:45:31 +0100
hi,
i had the same problem with suse 6.2,
a tar-file on cd with 50 MB was shown as 15 MB
and was corrupt.
but this problem didn't occure on
suse 6.4 and later.
hagen
Brian Harvey schrieb:
>
> Hi everyone,
>
> I recently wanted to "clone" a Linux installation to another machine.
> Typically, I've just plugged the new machine's drive into the original
> one and tarred it all over, but this time I wanted to put the original
> onto a cd for ease of cloning many machines.
>
> I tarred the original to file and copied to my winnt box and burned it
> onto a cd. Here's where the problem comes into the picture. I kept
> getting isofs_read_level3_size errors on the new Linux box, so I brought
> the cd back to the NT box and it was fine. Winzip opened the tarball -
> no problem. Hmmm......
>
> So I went back to the original Linux box and split the tarball into 1Meg
> pieces (350 of them), copied these to the NT box and burned another cd.
>
> Back to the new Linux box, I concatenated the split files together and
> it untarred just fine. So, there must be a problem with Linux reading
> very large files from a burned cd.
>
> If anyone can shed light on this issue, I'd be curious to know what's
> up. Has anyone out there seen this isofs_read_level3_size
> diagnostic error message?
>
> Thanks,
>
> -B
>
> --
> Brian Harvey
> Abacus Travel, Inc.
> 5 Lakeland Park Drive
> Peabody, MA 01960
> 978.326.3203 Voice
> 978.535.7770 Fax
> http://www.abacustravel.com
------------------------------
From: jens <[EMAIL PROTECTED]>
Subject: Re: ECC RAM supported by Linux?
Date: Sat, 09 Dec 2000 15:32:10 GMT
Thanks Chuck and Rico for some very interesting points. I think that
my next machine will in fact have ECC even if the failure rate is
extremely low - just for 'peace of mind'
Jens
------------------------------
From: walt <[EMAIL PROTECTED]>
Crossposted-To: linux.debian.user,comp.os.linux.setup
Subject: Re: Trying to port a hacked CMI8330 audio setup from RH6.2 to Potato...
Date: 09 Dec 2000 08:09:06 -0800
Michael V. Ferranti <[EMAIL PROTECTED]> writes:
> Hi guys,
>
> Originally, I used to have a CMI8330 HOWTO from a Slackware 2.1 setup
> that included a small C program to tweak the chip into working. I had it
> sloppily hacked into RH6.2's bootup scripts and it worked fairly well. I
> have the new HOWTO now that doesn't require the small C program, but it's
> written for RH6.1 and their /etc/rc.d/rc.sysinit file. Has anyone tried
> this with Debian Potato, or at least know RH6.1's order of events for
> /etc/rc.d/rc.sysinit? I'd like to insert the following in Debian Potato
> without violating any of the policies:
>
> ># Insert into /etc/conf.modules:
> >alias sound sb
> >options sb io=0x220 irq=5 dma=1 dma16=5
> >alias midi mpu401
> >options mpu401 io=0x330 irq=10
> >alias synth opl3
> >options opl3 io=0x388
The lines above definitely belong in /etc/modutils/aliases,
after which you type (as root) update-modules. This updates
the /etc/modules.conf file for you.
>
> ># Insert into /etc/rc.d/rc.sysinit at line 190:
> ># load sound modules
> >if [ -n "$USEMODULES" ]; then
> > if grep -s -q "^alias sound" /etc/conf.modules ; then
> > action "Loading sound module" modprobe sound
> > fi
> > if grep -s -q "^alias midi" /etc/conf.modules ; then
> > action "Loading midi module" modprobe midi
> > fi
> > if grep -s -q "^alias synth" /etc/conf.modules ; then
> > action "Loading synth module" modprobe synth
> > fi
> >fi
I'd say stick this stuff in a shell script in /etc/rc.boot,
where it won't be replaced with the next system update.
------------------------------
From: "Enrico Ng" <[EMAIL PROTECTED]>
Subject: fasttrak66
Date: Sat, 9 Dec 2000 10:27:56 -0600
I am having problems with using the promise fasttrak66 with slackware.
I have a fasttrak66 with two identical drives striped. I want to install
slackware (on a drive on my regular IDE) but the cd wont boot up all the
way.
when I try booting with ata66.i it gets to
PDC20262: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary
MASTER Mode.
ide2: ...
and just hangs.
booting off the cd doesnt even get to this step.
I was wondering if anyone ever got this to work?
thanx
--
Enrico Ng <[EMAIL PROTECTED]>
------------------------------
From: "Lenny Shovsky" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux
Subject: harddrive upgrade
Date: Sat, 9 Dec 2000 11:38:31 -0500
I ipgraded my ide system with ut2 scsi controller ( tekram-symbios) and 2
ibm u2w scsi hds. It detected the controller properly, but I still can not
connect to any of scsi hds. scsi ids are 0 and 1. The devices I tried to
access were /dev/sda and /dev/sdb
( fdisk /dev/sda). The error message I get is "Unable to open /dev/sda" .
When I run diskdrake it also shows only the ide drive I have in the system.
What do I need to do ?
thanx
------------------------------
From: David <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.admin,comp.os.linux.setup
Subject: Re: please help !!!! Format hd problem
Date: Sat, 09 Dec 2000 12:05:45 -0600
Fabien Guignon wrote:
>
> Hello all,
>
> here's my problem.
> I made an install of NT server and Linux red hat on one hd.
> My boss wants me to put this disk in a new pc with nt workstation only.
> is there a way to format all my hd in one time ?
>
> TIA
> Fabien
To remove all partitions and wipe the drive clean.
mknod /dev/hdX (where hdX or sdX is your drive)
cat /dev/zero >/dev/hdX
--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter. http://counter.li.org
ID # 123538
Completed more work units than 98.873% of seti users. +/- 0.01%
------------------------------
From: ds1436 <[EMAIL PROTECTED]>
Subject: Re: Running a headless box
Date: Sat, 09 Dec 2000 18:05:36 GMT
Look for the "halt on : no errors" option in your BIOS. If it's not there, and
there's no other settings for keyboards, there are some other things you can do.
I've seen instructions on the internet for making 'loopback' keyboard cables
(which trick the BIOS into thinking that there's a keyboard attached). I've also
heard of people removing the circuit board from their keyboards, and putting it
inside the computer case. (Newer keyboards have a small board, about 2" square,
that connects the keyboard cable to the flexible circuit sheet that receives the
keystrokes... open up a newer keyboard and you'll see what I mean.)
Mark Riehl wrote:
> All,
>
> I'd like to run a headless box (no monitor or keyboard) sitting on the floor
> for backups. It will be an older 486 with a LAN card and a video card.
>
> Is there a trick that will let me boot w/o a keyboard? Or, do I need to
> boot with the keyboard and then pull it off?
>
> Thanks,
> Mark
------------------------------
From: [EMAIL PROTECTED] (Ned Latham)
Crossposted-To: alt.os.linux
Subject: Re: harddrive upgrade
Date: 9 Dec 2000 18:28:35 GMT
Reply-To: [EMAIL PROTECTED]
Lenny Shovsky wrote in <[EMAIL PROTECTED]>:
>
> I ipgraded my ide system with ut2 scsi controller ( tekram-symbios) and 2
> ibm u2w scsi hds. It detected the controller properly, but I still can not
> connect to any of scsi hds. scsi ids are 0 and 1. The devices I tried to
> access were /dev/sda and /dev/sdb
> ( fdisk /dev/sda). The error message I get is "Unable to open /dev/sda" .
> When I run diskdrake it also shows only the ide drive I have in the system.
> What do I need to do ?
> thanx
I take that before the upgrade you had only IDE drives?
What kernel did you install? Was it bare.img or some such?
If so, you need to recompile for SCSI support, or reinstall.
You'll also need to install the right SCSI module for your card.
Also, check for interrupt conflicts.
--
Ned ++++++ Democracy means "the people rule".
Forget the republic: fight for the power of assent.
To reply, cut out my nose and make the met a net.
------------------------------
From: "eftech_1-=-" <[EMAIL PROTECTED]>
Subject: which ram?
Date: Sat, 09 Dec 2000 18:39:35 GMT
hi, i have a choice of generic pc133 ram or kingston 133 ram. is kingston
better? and will it really make a difference? just want to know if i should
spend the extra 20 bucks. thanks
------------------------------
From: [EMAIL PROTECTED] (Dances With Crows)
Subject: Re: ZIP Plus - need imm.tar.gz
Date: 9 Dec 2000 18:58:49 GMT
Reply-To: [EMAIL PROTECTED]
On Fri, 08 Dec 2000 12:29:55 -0600, Doug staggered into the Black Sun and said:
>I had my ZIP Plus (250M - Parallel Port) drive running under Linux when
>I first got it. Unfortunately, I since discarded the driver and have
>not been able to find it anywhere (imm.tar.gz/imm.o). Did Iomega put
>the clamps down? Can anyone provide me with the driver?
/usr/src/linux/drivers/scsi/imm.c ?
Do a "make menuconfig" and enable imm support. It's under "SCSI
Lowlevel drivers" in the SCSI Support section.
--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / Workin' in a code mine, hittin' Ctrl-Alt
http://www.brainbench.com / Workin' in a code mine, whoops!
=============================/ I hit a seg fault....
------------------------------
From: [EMAIL PROTECTED] (Carlos)
Subject: Re: ECC RAM supported by Linux?
Date: 9 Dec 2000 12:53:25 -0600
OK, I downloaded your program and this is the output I get from running
it:
440BX chipset detected
DRAM Data Integrity Mode = ECC with scrubbing
Error Command Register = 0x0002
Error Status Register 0 = 0x0000
Error Address Pointer Register = 0x03812000
What does it mean?? How many errors (if any) have been registered since
the last reboot (uptime shows: 12:52pm up 109 days, 2:10, 8 users,
load average: 0.07, 0.09, 0.07)?
Thanks!!
Carlos
please disregard email in header and use carlos1w @ excite.com
In article <90sesr$139$[EMAIL PROTECTED]>,
Rico Tudor <[EMAIL PROTECTED]> wrote:
>Linux, and any other OS, will benefit from ECC (Error Correcting Code) if
>
> (1) Your PC is outfitted with "ECC" main memory, and
> (2) The motherboard chipset has ECC circuitry, and
> (3) The circuitry is enabled by the BIOS
>
>Mainstream PC architecture defines a 72-bit wide path to main memory:
>64 for data and 8 for parity (formally, the ECC syndrome). Most people
>save money and purchase a 128-MByte DIMM like
>
> http://www.crucial.com/index.asp?imodule=CT16M64S4D8E
>
>This has 16 RAM chips in total, 8 per side, and costs $65.69 right now.
>If the chipset and BIOS support ECC, it must be disabled.
>
>If you bought a 128-MByte ECC DIMM like
>
> http://www.crucial.com/index.asp?imodule=CT16M72S4D8E
>
>you would see 18 chips, 9 per side, and would pay $87.29. If the chipset
>and BIOS support ECC, you would want to enable the ECC function.
>
>Memory failures can be caused by cosmic rays, alpha particles, a
>faulty RAM chip, bad power, or marginal signals around the system.
>Assuming an error in the data or syndrome bits, the ECC mathematical
>function used in 72-bit systems can detect all 1- and 2-bit errors.
>Further, the function can correct all 1-bit errors. Such correction is
>applied by the chipset "on the fly", meaning there is no CPU intervention.
>Intel logs this "soft error" within their chipsets, for later access by
>the OS. Since Linux doesn't bother, I wrote my own user-level utility:
>
> http://patrec.com/rico/ecc/
>
>On ECC error, chipsets like the 440BX can be programmed to assert signal
>SERR# (see "82443BX Host Bridge Datasheet", page 4-26). Motherboards
>will often route this to the south bridge to allow NMI interrupts. So,
>the famous Linux "Dazed and confused" message often means an ECC error.
>Regarding NMI enable, there are policy variations among motherboard
>vendors: Asus doesn't, Tyan does. My utility is unaffected, since it
>polls the chipset registers.
>
>Detection requires that the erroneous memory location be read. That means
>usage patterns imposed by the particular OS can alter when the chipset
>encounters the error.
>
>-------- my e-mail domain: math.nwu.edu
------------------------------
** 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
******************************
