Linux-Hardware Digest #100, Volume #13 Fri, 23 Jun 00 11:13:11 EDT
Contents:
Re: Is my Athlon going bad? ("B. Joshua Rosen")
Re: ** HP machines w/ integrated graphics ** (Rick Emery)
Re: Is my Athlon going bad? (Bernhard Mogens Ege)
Re: Help on motherboard decision ("Tom Brinkman")
Modem-based answering machine (Wendell Craig)
Re: Is my Athlon going bad? (Martin Herrman)
D-Link DU-H4SMK USB Hub and Redhat 6.2 (NikosKalogridis)
dhcpd memory hog? ([EMAIL PROTECTED])
Re: Getting RH Linux to recognize ATA66 IDE controller ("Charles Sullivan")
----------------------------------------------------------------------------
From: "B. Joshua Rosen" <[EMAIL PROTECTED]>
Subject: Re: Is my Athlon going bad?
Date: Fri, 23 Jun 2000 09:20:26 -0400
Several questions,
1) Did your system become unstable after a software upgrade, or did it
just happen?
2) Have you added any new hardware?
3) You mentioned that you were running Netscape when it hung, is it
always Netscape that's running when it hangs? Netscape is a piece of
crap that crashes all the time, and frequently wedges X Windows in the
process. Ctrl Alt Backspace usually can fix it, but not always. Netscape
is very sensitive to communications problems so if your internet
connection has become less reliable then you'll see more Netscape
problems.
A few things to try,
It's possible that you have a hardware problem, but the first thing to
do is make sure all of your DIMMS, boards and cables are seated
properly. Open up your machine and jiggle everything.
It's also possible that something in the OS has been corrupted or that
you added something that was incompatiable. I'd grab an up to date
distribution, like RedHat 6.2, and do a clean install. By clean I mean
reformat the / partition and then re-install the OS. An upgrade is
likely to make things worse and won't fix any problems that are caused
by a bad settings. A clean install will get you a consistant set of
components and will get all of the configurations back to a known good
baseline. If you do this make sure all of your data is on a different
partition, and copy /etc over to a different partition so that you can
reference your old settings (mostly fstab and the various networking
configurations) when you are reconfiguring the system after the
re-install.
Finally, hitting the reset button is what caused the file system problem
that resulted in the manual fsck, not the prior crash. However a
filesystem problem severe enough to require a manual fsck means that you
might have lost something important, another reason to do a clean
install of the OS.
Bernhard Mogens Ege wrote:
>
> My Athlon is freezing more often than I care for. It has happened 4
> times now, 3 times at night, and 1 just now. This last time it hasn't
> been running more than 16 hours!
>
> I am using the 2.2.16 kernel with the ide driver patch applied (to
> make it recognize my ide controller). The patch I do not suspect as
> the crashing has occured before I used that patch.
>
> The crash was as follows:
>
> I watched netscape draw a page (loading big image from net) and the
> machine suddenly stopped responding (no mouse, keyboard). From another
> machine I could ping it (kernel responed normally), and a telnet and
> ftp did connect, but the daemons never got further than establishing
> the connection (no HD access).
>
> ctrl-alt-del did not work. ctrl-alt-backspace neither. alt-sysreq
> isn't compiled in.
>
> Nothing else to do than push hardware reset (really hate that part).
>
> On bootup, fsck did not detect any errors on the filesystem
> (fortunately! but on earlier crashes a manual fsck was required). Just
> as I was logged in (in X) the HD spun down and then spun up again!
> This I have never experienced before.
>
> What can be wrong with my system?
>
> The first 6 month of Linux were without crashes, now they suddenly
> turn up. I need this machine to work at all times (always on).
>
> regards,
>
> Bernhard Ege
>
> System: Athlon Classic 500, MSI 6167, Western Digital Expert 18Gb drive, 128Mb RAM,
>50x CDROM, 3com network interface (100Mbit)
------------------------------
From: Rick Emery <[EMAIL PROTECTED]>
Subject: Re: ** HP machines w/ integrated graphics **
Date: Fri, 23 Jun 2000 09:20:39 -0400
Carlos wrote:
> Hi, does anyone know if the cheaper HP machines with "integrated 2d/3d
> graphics" work with XFree86? I know they are not top performance machiens
> but I don't care about that...
>
> Thank you!
>
> Carlos
>
> Please respond to [EMAIL PROTECTED] (not to the address in the
> header!!).
I have a Pavilion 8560C with the Intel i810 chipset. Support for this 2d/3d
chip is not built into XFree86, but support.intel.com has Linux drivers for
it and specific instructions for installing it. I found it to be quite easy.
Hope this helps.
------------------------------
Subject: Re: Is my Athlon going bad?
From: Bernhard Mogens Ege <[EMAIL PROTECTED]>
Date: Fri, 23 Jun 2000 13:42:38 GMT
>>>>> "B" == B Joshua Rosen <[EMAIL PROTECTED]> writes:
> Several questions,
> 1) Did your system become unstable after a software upgrade, or did it
> just happen?
Just happened as far as I can tell (the first anyway, since then I
have upgraded kernel and orther packages).
> 2) Have you added any new hardware?
No, no hardware changes at all.
> 3) You mentioned that you were running Netscape when it hung, is it
> always Netscape that's running when it hangs? Netscape is a piece of
> crap that crashes all the time, and frequently wedges X Windows in the
> process. Ctrl Alt Backspace usually can fix it, but not always. Netscape
> is very sensitive to communications problems so if your internet
> connection has become less reliable then you'll see more Netscape
> problems.
Keyboard was dead, so was the mouse. Couldn't telnet or ftp in (the
sockect connection was established, though). The first crash caused a
complete kernel lockup (no response) and hardware reset was my only
option.
Can netscape cause a complete system crash (no telnet access)? I can
understand netscape killing X (has happened some times to me) but the
kernel?
> A few things to try,
> It's possible that you have a hardware problem, but the first thing to
> do is make sure all of your DIMMS, boards and cables are seated
> properly. Open up your machine and jiggle everything.
Good ide, completely forgot that one. This could help. Maybe use less
aggressive timings in the bios (haven't toughed them for several
month, though).
> It's also possible that something in the OS has been corrupted or that
> you added something that was incompatiable. I'd grab an up to date
> distribution, like RedHat 6.2, and do a clean install. By clean I mean
> reformat the / partition and then re-install the OS. An upgrade is
> likely to make things worse and won't fix any problems that are caused
> by a bad settings. A clean install will get you a consistant set of
> components and will get all of the configurations back to a known good
> baseline. If you do this make sure all of your data is on a different
> partition, and copy /etc over to a different partition so that you can
> reference your old settings (mostly fstab and the various networking
> configurations) when you are reconfiguring the system after the
> re-install.
Hmmm, this I would rather avoid as I have about 3.5Gb stuff laying
around (including the rh60 install). If I reduced the extra modules
and disabled NFS export and stuff like that, maybe I could locate the
cause.
> Finally, hitting the reset button is what caused the file system problem
> that resulted in the manual fsck, not the prior crash. However a
> filesystem problem severe enough to require a manual fsck means that you
> might have lost something important, another reason to do a clean
> install of the OS.
Why do you think the reset caused the file system problem and not the
kernel crash? For all I know, the ide driver could have been halted
during disc update.
I just wish that Linux was able to store crash data somewhere on the
disc (or on floppy) when a crash occured.
The log just doesn't mention anything about the crash...
Jun 23 12:45:33 overmind PAM_pwdb[1155]: (su) session closed for user root
Jun 23 12:45:45 overmind PAM_pwdb[1232]: (su) session opened for user root by
bme(uid=21359)
Jun 23 13:18:55 overmind syslogd 1.3-3: restart.
Jun 23 13:18:55 overmind syslog: syslogd startup succeeded
Jun 23 13:18:55 overmind syslog: klogd startup succeeded
Jun 23 13:18:55 overmind kernel: klogd 1.3-3, log source = /proc/kmsg started.
Crash happened between 12:45 and 13:18...
regards,
Bernhard Ege
> Bernhard Mogens Ege wrote:
>>
>> My Athlon is freezing more often than I care for. It has happened 4
>> times now, 3 times at night, and 1 just now. This last time it hasn't
>> been running more than 16 hours!
>>
>> I am using the 2.2.16 kernel with the ide driver patch applied (to
>> make it recognize my ide controller). The patch I do not suspect as
>> the crashing has occured before I used that patch.
>>
>> The crash was as follows:
>>
>> I watched netscape draw a page (loading big image from net) and the
>> machine suddenly stopped responding (no mouse, keyboard). From another
>> machine I could ping it (kernel responed normally), and a telnet and
>> ftp did connect, but the daemons never got further than establishing
>> the connection (no HD access).
>>
>> ctrl-alt-del did not work. ctrl-alt-backspace neither. alt-sysreq
>> isn't compiled in.
>>
>> Nothing else to do than push hardware reset (really hate that part).
>>
>> On bootup, fsck did not detect any errors on the filesystem
>> (fortunately! but on earlier crashes a manual fsck was required). Just
>> as I was logged in (in X) the HD spun down and then spun up again!
>> This I have never experienced before.
>>
>> What can be wrong with my system?
>>
>> The first 6 month of Linux were without crashes, now they suddenly
>> turn up. I need this machine to work at all times (always on).
>>
>> regards,
>>
>> Bernhard Ege
>>
>> System: Athlon Classic 500, MSI 6167, Western Digital Expert 18Gb drive, 128Mb RAM,
>50x CDROM, 3com network interface (100Mbit)
--
Bernhard Mogens Ege, M.Sc.E.E. E-mail: [EMAIL PROTECTED]
Medical Informatics and Image Analysis Direct call: +45 96 35 87 82
Institute of Electronic Systems Switchboard: +45 96 35 80 80
Aalborg University Fax: +45 98 15 40 08
Frederik Bajersvej 7, D1-203 Homepage:
DK-9220 Aalborg East http://www.vision.auc.dk/~bme
==============================================================================
Home: Hadsund Landevej 454, DK-9260 Gistrup, Phone: +45 96365086, +45 22749713
------------------------------
From: "Tom Brinkman" <[EMAIL PROTECTED]>
Subject: Re: Help on motherboard decision
Crossposted-To: alt.comp.hardware.homebuilt
Date: Fri, 23 Jun 2000 09:03:26 -0600
In article <8iulm5$7fd$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>>
>> Asus or Soyo. Get a 700e or a 750e, but not a 733. You'll
>> have the best system with a BX board and a 100mhz FSB p3, ie, an
>> 'e', but not an 'eb'. Use at least pc100 cas2 ram (8ns, CL2)
>> non ECC
>>
>> The i8<whatever> chipsets are even worse than you're read,
>> VIA's aren't much better.... go BX. I like the Soyo 6ba+IV
>> --
>> ~~ Tom Brinkman [EMAIL PROTECTED]
>
> Is there any BX FCPGA boards you can recommend, or is it better
> off go with a slot 1? How about ASUS CUBX? Is there any advantage
> to get a 133MHz board for 700e or 750e for possible overclocking?
>
> Thanks, Xiaobin
Not really, I've always used slot1 BX, but there's a lot of hands on
alt.comp.hardware.overclocking with good BX-flip chip advice. There's
a lot of kids there with wild ideas, so give more attention to the regular
posters, specially those that give links to websites like Tom's Hardware
or Anand (both of which concluded like me that the best system right
now is still BX based)
There's 2 very good reasons for NOT getting a 133mhz board, or 133mhz
cpu. 1) if it's 133mhz, it's not BX. That said, a BX board will run
cpu's at 133mhz and higher (155mhz!) with less problems and better
performance than the 133mhz chipset boards, eg, VIA and i8<whatever>
2) revolves around AGP and BX. A BX board will do a better, less
problems, better stabilitly job of running an AGP card, but not at
FSB speeds much over 112mhz for most cards. The reason is BX doesn't
support (and doesn't introduce the problems) of running asychronous
bus speeds, and doesn't support AGP dividers other than 1:1 and 1:2/3.
It will support higher PCI dividers at bus speeds at or over 100mhz tho.
What this means is, for example, with a p3-700e, the front side
bus (FSB) is the normal 100mhz (cpu is at 7x100), the PCI bus speeds
are normal at 100/3 = 33.3mhz, and the AGP bus is normal at 100*2/3=
66.6mhz. Everythings right on spec.
It can also mean versatility with stability. My system is a good
example. P3-450, but run on a BX board at 135mhz! Cpu is at 4.5x135
(608mhz), PCI is 135/4=33.7 (close as PCI ever gets to right on spec).
AGP is 135*2/3= 90mhz. Since there's very few cards that will tolerate
an AGP bus that far off-spec, I use a PCI Voodoo3 .... no problems.
I suspect the reason 3DFX is continuing to produce PCI cards, eg, the
new 5500 Voodoo, is to accomodate people like me (an they also know
that VIA and i8xx boards suck)
You take a p3-700e, put it on a BX Asus or Soyo, run it at 7x133
(933mhz!) with some pc133 7ns cas2 ram, and a pci Voodoo55 and you'll
have somethin :->>
--
~~ Tom Brinkman [EMAIL PROTECTED]
------------------------------
From: Wendell Craig <[EMAIL PROTECTED]>
Subject: Modem-based answering machine
Date: Fri, 23 Jun 2000 14:18:34 GMT
For the third time in as many years, I'm going to tackle this
problem. I want to use my Linux box (Pent 550, 128 meg RAM, RedHat
6.2) as a 'phone answering system. I have a no-name voice-fax
external modem which I believe 'should' work for this purpose.
In previous tries - with lesser equipment - I've gotten nowhere, but
it seems to me I now have the horsepower and OS muscle to get such a
system working.
Would someone be willing to point me in a proper first direction? I'd
be grateful for any success stories and/or horror stories as well.
Thanks.
--
Wendell Craig - [EMAIL PROTECTED]
Ever notice that all the recent bugs only affect
computers running Windows? Ever wonder why?
------------------------------
From: [EMAIL PROTECTED] (Martin Herrman)
Subject: Re: Is my Athlon going bad?
Reply-To: [EMAIL PROTECTED]
Date: 23 Jun 2000 14:22:23 GMT
On Fri, 23 Jun 2000 13:42:38 GMT, Bernhard Mogens Ege <[EMAIL PROTECTED]>
wrote:
>
> Can netscape cause a complete system crash (no telnet access)? I can
> understand netscape killing X (has happened some times to me) but the
> kernel?
>
> The log just doesn't mention anything about the crash...
>
> Jun 23 12:45:33 overmind PAM_pwdb[1155]: (su) session closed for user root
> Jun 23 12:45:45 overmind PAM_pwdb[1232]: (su) session opened for user root by
bme(uid=21359)
> Jun 23 13:18:55 overmind syslogd 1.3-3: restart.
> Jun 23 13:18:55 overmind syslog: syslogd startup succeeded
> Jun 23 13:18:55 overmind syslog: klogd startup succeeded
> Jun 23 13:18:55 overmind kernel: klogd 1.3-3, log source = /proc/kmsg started.
>
> Crash happened between 12:45 and 13:18...
Netscape can't crash the kernel, but it can use 99.9 % of the system recourses,
then it will look like a crash, but the system will recover one day. The
problem
is: which day? ;-) So it wasn't a real crash. Some won't call this a crash,
I will. For me, the definition of a crash is: when the system doesn't respond
within 10 minutes.
You could use 'ulimit' to tell the kernel that Netscape isn't allowed to use
more than for example 90% of the recourses, you then will always be able to
kill netscape.
Martin
P.S. tell me if i'm not right please.
--
Linux Gebruikers Handleiding v1.2 : http://2mypage.cjb.net
Linux RedHat 6.1 Kernel 2.2.14 Toshiba P233 MHz, 32 Mb RAM
4:10pm up 12 days, 12:00, 2 users, load average: 0.09, 0.07, 0.01
Western Civilization, that would be a good idea!
------------------------------
From: [EMAIL PROTECTED] (NikosKalogridis)
Subject: D-Link DU-H4SMK USB Hub and Redhat 6.2
Date: 23 Jun 2000 14:28:55 GMT
Does anyone has this hub working on a redhat 6.1/6.2 box? I am thinking of
buying one to release the keyboard,ps/2 mouse and com port interrupts?
Any comment would be very much apprecieted?
Thanks,
Nikos Kalogridis,
email: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: dhcpd memory hog?
Date: Fri, 23 Jun 2000 14:15:45 GMT
I use dhcpd on a couple of machines. On all machines but one, dhcpd
uses about 500 to 700 K of memory. On the other machine it occupies
8000 to 9000 K. It does servers about 25 machines right know but even
with no leases it used the same memory.
Also, the memory is not grow like a mem leak. It starts at 8000 to
9000 k and stays there.
This is odd right. Any ideas?
The machines are RedHat 6.1 and 6.2. 2.2.16. Same version of dhcpd.
Thanks.
[EMAIL PROTECTED]
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Charles Sullivan" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Getting RH Linux to recognize ATA66 IDE controller
Date: Fri, 23 Jun 2000 14:31:18 GMT
Charlie Zender <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Hi,
>
> Redhat linux 6.2 will not install on my machine, it quits installation
> procedure with the error "No valid devices were found on which to create
> new filesystems". I have it from a helpful netizen that the problem is
> that lilo does not know how to find the ATA66 controller, and the
suggested
> fix is booting with "linux ide2=0x1080" and eventually
> adding a line to lilo.conf like so:
>
> >The bottom line is that Linux cannot find the ATA 66.
> >Here's what the core of my lilo.conf eventually looked
> >like after I got things installed:
>
> >image=/boot/vmlinuz-2.2.14-5.0
> > label=linux
> > append="ide2=0x1080"
> > read-only
> > root=/dev/hde5
>
> >other=/dev/hde1
> > label=dos
>
> Before I try this solution I have some questions:
> 1. Is this solution safe? Is it alway ide2? How do I know whether
> to use ide1 instead? What if I'm wrong with that hex address? Will
> it trash my windows partition?
>
> 2. Is this a generic problem with ATA66 controllers? with just
> the Promise ATA66 controllers? Is the problem fixed in kernel 2.4.0.test?
> Or should I just assume the problem will still be in 2.4 and install
> Redhat 6.2 now?
>
> 3. Is there a better solution? e.g., making a driver disk with the
> "right" ata66 module on it?
>
> Thanks,
> Charlie
>
> Here is my hardware description
> > Dell Dimension T Minitower 800 MHz PIII
> > 256 MB RAM
> > 40 GB Ultra ATA 7200 RPM with ATA 66 controller
> > NIC: 3Com EtherLink 10/100 PCI PCI For Complete PC Management
(3c905c-TX)
> > NVIDIA TNT2 M64 4xAGP w 32MB RAM
> > Microsoft PS/2 Mouse (Intellipoint)
> > Promise Technology Inc. Ultra66 IDE Controller
> > Intel 82371AB/EB PCI Bus Master IDE Controller
> > 120 MB Super disk, aka LS120
> > Sony 8X/4X/32X CD-RW drive
>
> --
> Charlie Zender [EMAIL PROTECTED] (949) 824-2987/FAX-3256, Department of
> Earth System Science, University of California, Irvine CA 92697-3100
Ide0 and ide1 are the ide ports on your motherboard. Ide2 is the primary
port on your add-on card. Since you have Windows installed, you can get the
correct hex code to use from Windows.
Do this: Click on Start/Settings/Control Panel/System/Device Manager.
Expand "SCSI controller". Highlight "Promise ... Ultra66 IDE Controller",
click on "Properties" and select tab "Resources". Look for the _very first_
entry for "Input/Output Range" (there are likely to be several of them)
and use the lower address (which is hex).
E.g., my system has "Input/Output Range 1440-1447" so I used "ide2=0x1440"
for installation and subsequently as the append in my /etc/lilo.conf file.
The RH 6.2 installation proceeded very smoothly. The partitions on the HD
attached
to the Ultra 66 card are devices /dev/hde1, hde2, etc.
Note: I don't know of a way to add this to a Linux boot floppy, so as soon
as possible I copied the Linux kernel to my Win 98 partition so I could use
LOADLIN.EXE to boot Linux in case my MBE were to ever get corrupted.
(If anyone knows how to create a Linux boot floppy which will work with
the HD on the Ultra66 card, I'd like to hear about it.)
------------------------------
** 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
******************************
