Linux-Hardware Digest #619, Volume #14 Fri, 13 Apr 01 13:13:06 EDT
Contents:
Neuhaus Smarty 28.8 TI not sending faxes, only receiving (Hans Posthumus)
Re: Random tape device... (Joshua Baker-LePain)
Sandisk SDDR31 USB Help needed (Schubert)
Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!! (Anonymous)
Re: Audio with Linux (2.4.2/2.2.16) kernel and Asus AV7133 ("Unstable")
Re: Samba - Strange Password Prompt (HOT/URGENT!) (Mark_Harju)
Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!! (Anonymous)
Re: Random tape device... (John Ouellette)
Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!! ("Peter T. Breuer")
scsi module aha1542.o doesn't load ("Rainer Traut")
Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!! (Anonymous)
Any thoughts on a good Sparc Linux? ("User Rdkeys Robert D. Keys")
Re: today's harddrives will surely fail before dialup users manage to fill them up?
("Troy Bowman")
Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!! ("Peter T. Breuer")
Re: Random tape device... (Joshua Baker-LePain)
----------------------------------------------------------------------------
Date: Fri, 13 Apr 2001 15:37:55 +0200
From: Hans Posthumus <[EMAIL PROTECTED]>
Subject: Neuhaus Smarty 28.8 TI not sending faxes, only receiving
Hi,
perhaps someone of you is more succesful in using this faxmodem under
hylafax ;-)
With my current config.modem file receiving faxes is no problem at all.
But when sending a fax the modem starts connecting and then it seems not
finding the right "handshake" with the other faxmachine.
Anyone having functioning config.modem parameters for this modem?
Bye
Hans Posthumus
------------------------------
From: Joshua Baker-LePain <[EMAIL PROTECTED]>
Subject: Re: Random tape device...
Date: 13 Apr 2001 13:45:03 GMT
John Ouellette <[EMAIL PROTECTED]> wrote:
> on hand and aren't really used for writing. The DLT, on the other hand,
> is used heavily by users and for system backups. The problem is that
> *sometimes* the DDS drive isn't detected during boot-up -- this results
I would try to look at *why* this happens, as it seems indicitave of
some hardware issues, and fixing them would be good in a number of ways.
Check the usual suspects -- cable length, termination, age/color of goat.
Obviously, make sure that all the drives are powered up when you boot.
> in the DLT drive changing it's device number (i.e. becomes /dev/nst1
> instead
> of /dev/nst2). Is there a way of forcing a tape drive to always be the
> same
> tape device?
> Now that I'm writing this, I'm not 100% certain what order the SCSI bios
> is looking at the interfaces: the Exabyte and DDS on a PCI Adaptec 2930
> SCSI card while the DLT is on an on-board Ultra2 interface (the disks
> are
> on a second on-board channel). Hmmmm.... I may have partially answered
Hmmm, indeed. If the DLT is being assigned to nst2, that means that
it's bus is being probed after the 2930 bus. What driver do the onboard
channels use?
> my own question. However, *is* there away, regardless of the order
> in which the SCSI devices are probed by the bios, to ensure that my DLT
> will always be /dev/nstX?
Well, with 2.2.x, not really. You *could* try to make sure that the busses
get probed in the order you want. Ideally, the onboard bus with the DLT
on it would get probed before the 2930 -- that way your DLT would always be
/dev/nst0, regardless of whether or not the DDS gets detected. I recall
editing some file in the kernel sources before to do this so that an aha152x
got probed *after* the onboard aic7xxx, but I don't remember which file it
was (and that was a while ago). Maybe turning the BIOS off on the 2930
would help?
With 2.4.x, you have the option of using devfs, which refers to devices
in a more heirarchical structure, i.e. (this is not exactly it, but
illustrative) /dev/scsi0/id0/lun0. Obviously, then, devices will always
be where you expect them to be.
Good luck.
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
------------------------------
From: Schubert <[EMAIL PROTECTED]>
Subject: Sandisk SDDR31 USB Help needed
Date: Fri, 13 Apr 2001 16:13:13 +0200
I want to access my compactflash memory card via the sandisk SDDR31
Reader (USB) to the linux system.
I'm using SuSE Distr.+ Kernel 2.2.16 and compiled the USB-modules
(Besemodule and the Mass Storage Module).
>From the Internet I know that the SDDR31 is supported. I don't know if
it is running with the mass storage module that came with the
distribution.
Something is wrong. Basiscally the usb-port is recognized( see following
excerpt from /var/log/messages)
linux kernel: usb.c: USB new device connect, assigned device number 2
linux kernel: usb.c: This device is not recognized by any installed USB
driver.
linux kernel: hub.c: already running port 2 disabled by hub (EMI?),
re-enabling...
linux kernel: usb.c: USB disconnect on device 2
linux kernel: usb.c: USB new device connect, assigned device number 2
linux kernel: usb.c: This device is not recognized by any installed USB
driver.asfdasf
Is there a howto available that solves the problem or do you hve some
advice.
Thanks in advance
Christian
------------------------------
Date: Fri, 13 Apr 2001 09:33:47 -0500
Crossposted-To: comp.os.linux.misc
Subject: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!!
From: Anonymous <[EMAIL PROTECTED]>
Hi all,
I would really appreciate some assistance and follow up. Configuration
details are below. And my thanks in advance to all that reply.
We have several new Supermicro 6010 L\H series servers. They work fine
until you add more than 2 DIMMs ram. Then for the life of us we can not
keep them running. See the set up information below. They are more
stable under a 2.2 kernel than a 2.4 kernel but not entirely so. Also,
we run Netfilter and MySQL so being able to take advantage of the 2.4
kernel is a priority.
Once installed they will not compile anything without seg faults and
core dumps. They kernel panics about every 4 to 8 hours and we have
been unable to figure out what is going on.
We get this on different machines and we have tried lots of processor,
ram and drive substitutions.
I have made sure that APM is on in the BIOS and CAS is set to 3. I have
been administrating linux for 3 years now but only on single
processor/IDE machines. So I am at a complete loss.
Is there anything else I might be missing? Distro, Ram, BIOS???
Sincerely,
Scott ES
Configurations are:
Supermicro 6010 L\H (H is a dual channel scsi, L is single)
L
<http://www.supermicro.com/PRODUCT/SUPERServer/SUPER%20SERVER%206010L%20Server.htm>
H
<http://www.supermicro.com/PRODUCT/SUPERServer/SUPER%20SERVER%206010%20Server.htm>
BIOS set to Optimal Settings (APM must be on or it will not install)
Hardware additions: (all items were bought at the same time from one
vendor)
SMP Intel 833's running at 133mhz fsb
Reg ECC 128's or 256's (Micron chips on 3rd party silicon)
IBM 9GB, 18GB, and 36GB SCA Ultra 160's
Kernel's have been compiled with variations of the following:
Monolithic and Modulated.
Kernels 2.4.2, 2.4.3, 2.4.3-ac4, 2.4.4-pre1
SMP
MTRR support on and off
4GB memory support
SererWorks IDE Chipset
USB on and off
APM on and off
AIC7XXX drivers (old and new including the latest 6.1.11)
With and without Luns
5000 ms Initial bus reset delay
Netfilter on and off
QOS on and off
Compilers used:
kgcc-1.1.2-40
gcc-2.96-69
glibc:
compat-glibc-6.2-2.1.3.2
glibc-common-2.2-12
glibc-devel-2.2-12
glibc-2.2-12
glibc-profile-2.2-12
All other patches are current. (libstdc++, etc)
--------== Posted Anonymously via Newsfeeds.Com ==-------
Featuring the worlds only Anonymous Usenet Server
-----------== http://www.newsfeeds.com ==----------
------------------------------
From: "Unstable" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,alt.comp.periphs.mainboard.asus
Subject: Re: Audio with Linux (2.4.2/2.2.16) kernel and Asus AV7133
Date: Fri, 13 Apr 2001 10:36:55 -0400
It might be a configuration problem with the real player, not the onboard
sound.
Look at the settings in real player and see if something is set wrong.
"Larry Snyder" <[EMAIL PROTECTED]> wrote in message
news:%LrB6.6586$[EMAIL PROTECTED]...
> I have been unable to get the onboard audio to work with my Asua A7V133
> under Linux - it works just fine under Windows 2K.
>
> The audio hardware is detected at boot time - when I try to stream audio
> using the real player I get lots of static (but I can make out the music
in
> the background)..
>
> When I log into the machine (using Mandrake 8 Beta 3) I do get the "login"
> WAV played just fine..
>
> Any ideas?
>
> Larry
>
>
------------------------------
Crossposted-To: comp.os.linux.networking,comp.os.linux.setup
From: Mark_Harju <[EMAIL PROTECTED]>
Subject: Re: Samba - Strange Password Prompt (HOT/URGENT!)
Date: Fri, 13 Apr 2001 14:12:27 GMT
Many thanks to all who gave tips!
Interesting note: My Win95 laptop logs in to Samba without a problem, so
I'm guessing this issue is due to a Win98 password encryption feature
which must be disabled in the registry. A PhD from Oz tipped me off to
this. No worries, mate!
Thanks again to all who replied!
--
Opinions expressed are not
necessarily those of the
Boeing Company. Please
remove "NO-SPAM" to
reply. Thanks!
Mark_Harju wrote:
>
> Hello everyone. Sorry to bug you, but I've got a Linux hair-puller here!
>
> I'm at my wits' end trying to connect to my Samba Linux box via a win98
> machine, but every time I do, I'm getting a password prompt. No password
> I've got will work! I hope this isn't a dummy thing...!
>
> Here are the particulars via RTFM (Mandrake 6.0)
>
> 1. smb.conf file modified for sharing and correct workgroup name (same
> as the Win98 workgroup name, incl case); therefore -
> 2. visibility of Linux system in the Win box's "Network Neighborhood" is
> present
> 3. user account created in Linux
> 4. samba installed and active
> 5. can access server via "nobody" logon (though access is severely
> limited, the connection does get made, and displays folders on the Linux
> box)
> 6. Logging into the Win box via MS Client, using the userID and password
> established on the Linux box
> 7. Can ping either box from the other
> 8. Can access the Linux box from the Win98 box via Telnet.
> 9. I don't have it in front of me, but I'm pretty sure the Win98 box is
> setup for a workgroup, not a domain (this is the default)
>
> I've checked and checked my Linux settings, and everything looks
> correct. Can anybody please advise me as to what I may be overlooking? I
> need an answer by tomorrow (Friday)
>
> Please respond to [EMAIL PROTECTED] and my Boeing address.
> Please remove the "NO-SPAM-".
>
> Thanks much in advance. If you're in the Seattle area and you can help
> me, I'll pay you with a sixpack of Obsidian Stout!
>
> mh
------------------------------
Date: Fri, 13 Apr 2001 09:45:35 -0500
Crossposted-To: comp.os.linux.misc
Subject: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!!
From: Anonymous <[EMAIL PROTECTED]>
Hi all,
I would really appreciate some assistance and follow up. Configuration
details are below. And my thanks in advance to all that reply.
We have several new Supermicro 6010 L\H series servers. They work fine
until you add more than 2 DIMMs ram. Then for the life of us we can not
keep them running. See the set up information below. They are more
stable under a 2.2 kernel than a 2.4 kernel but not entirely so. Also,
we run Netfilter and MySQL so being able to take advantage of the 2.4
kernel is a priority.
Once installed they will not compile anything without seg faults and
core dumps. They kernel panics about every 4 to 8 hours and we have
been unable to figure out what is going on.
We get this on different machines and we have tried lots of processor,
ram and drive substitutions.
I have made sure that APM is on in the BIOS and CAS is set to 3. I have
been administrating linux for 3 years now but only on single
processor/IDE machines. So I am at a complete loss.
Is there anything else I might be missing? Distro, Ram, BIOS???
Sincerely,
Scott ES
Configurations are:
Supermicro 6010 L\H (H is a dual channel scsi, L is single)
L
<http://www.supermicro.com/PRODUCT/SUPERServer/SUPER%20SERVER%206010L%20Server.htm>
H
<http://www.supermicro.com/PRODUCT/SUPERServer/SUPER%20SERVER%206010%20Server.htm>
BIOS set to Optimal Settings (APM must be on or it will not install)
Hardware additions: (all items were bought at the same time from one
vendor)
SMP Intel 833's running at 133mhz fsb
Reg ECC 128's or 256's (Micron chips on 3rd party silicon)
IBM 9GB, 18GB, and 36GB SCA Ultra 160's
Kernel's have been compiled with variations of the following:
Monolithic and Modulated.
Kernels 2.4.2, 2.4.3, 2.4.3-ac4, 2.4.4-pre1
SMP
MTRR support on and off
4GB memory support
SererWorks IDE Chipset
USB on and off
APM on and off
AIC7XXX drivers (old and new including the latest 6.1.11)
With and without Luns
5000 ms Initial bus reset delay
Netfilter on and off
QOS on and off
Compilers used:
kgcc-1.1.2-40
gcc-2.96-69
glibc:
compat-glibc-6.2-2.1.3.2
glibc-common-2.2-12
glibc-devel-2.2-12
glibc-2.2-12
glibc-profile-2.2-12
All other patches are current. (libstdc++, etc)
--------== Posted Anonymously via Newsfeeds.Com ==-------
Featuring the worlds only Anonymous Usenet Server
-----------== http://www.newsfeeds.com ==----------
------------------------------
From: John Ouellette <[EMAIL PROTECTED]>
Subject: Re: Random tape device...
Date: Fri, 13 Apr 2001 11:14:43 -0400
Reply-To: [EMAIL PROTECTED]
How stable is devfs? I haven't thought about it in a while, but last
I remember there were 'issues'.... I'll be upgrading to 2.4.x in
the near future, so this might be an option.
Thx,
J.
Joshua Baker-LePain wrote:
>
> John Ouellette <[EMAIL PROTECTED]> wrote:
>
> > on hand and aren't really used for writing. The DLT, on the other hand,
> > is used heavily by users and for system backups. The problem is that
> > *sometimes* the DDS drive isn't detected during boot-up -- this results
>
> I would try to look at *why* this happens, as it seems indicitave of
> some hardware issues, and fixing them would be good in a number of ways.
> Check the usual suspects -- cable length, termination, age/color of goat.
> Obviously, make sure that all the drives are powered up when you boot.
>
> > in the DLT drive changing it's device number (i.e. becomes /dev/nst1
> > instead
> > of /dev/nst2). Is there a way of forcing a tape drive to always be the
> > same
> > tape device?
>
> > Now that I'm writing this, I'm not 100% certain what order the SCSI bios
> > is looking at the interfaces: the Exabyte and DDS on a PCI Adaptec 2930
> > SCSI card while the DLT is on an on-board Ultra2 interface (the disks
> > are
> > on a second on-board channel). Hmmmm.... I may have partially answered
>
> Hmmm, indeed. If the DLT is being assigned to nst2, that means that
> it's bus is being probed after the 2930 bus. What driver do the onboard
> channels use?
>
> > my own question. However, *is* there away, regardless of the order
> > in which the SCSI devices are probed by the bios, to ensure that my DLT
> > will always be /dev/nstX?
>
> Well, with 2.2.x, not really. You *could* try to make sure that the busses
> get probed in the order you want. Ideally, the onboard bus with the DLT
> on it would get probed before the 2930 -- that way your DLT would always be
> /dev/nst0, regardless of whether or not the DDS gets detected. I recall
> editing some file in the kernel sources before to do this so that an aha152x
> got probed *after* the onboard aic7xxx, but I don't remember which file it
> was (and that was a while ago). Maybe turning the BIOS off on the 2930
> would help?
>
> With 2.4.x, you have the option of using devfs, which refers to devices
> in a more heirarchical structure, i.e. (this is not exactly it, but
> illustrative) /dev/scsi0/id0/lun0. Obviously, then, devices will always
> be where you expect them to be.
>
> Good luck.
>
> --
> Joshua Baker-LePain
> Department of Biomedical Engineering
> Duke University
--
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
John Ouellette | Ph: 212-313-7919
Department of Astrophysics | Fax: 212-769-5007
American Museum of Natural History | e-mail: [EMAIL PROTECTED]
Central Park West at 79th St. |
http://research.amnh.org/astrophysics
New York, NY 10024-5192 |
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!!
Date: Fri, 13 Apr 2001 16:56:35 +0200
In comp.os.linux.misc Anonymous <[EMAIL PROTECTED]> wrote:
> We have several new Supermicro 6010 L\H series servers. They work fine
> until you add more than 2 DIMMs ram. Then for the life of us we can not
> keep them running. See the set up information below. They are more
> Once installed they will not compile anything without seg faults and
> core dumps. They kernel panics about every 4 to 8 hours and we have
Set the bios to very conservative values .. and check that the mobo
accepts the type of ram you have.
> I have made sure that APM is on in the BIOS and CAS is set to 3. I have
ON! You mean OFF, don't you? You can't have APM in an SMP machine.
> Is there anything else I might be missing? Distro, Ram, BIOS???
It's purely a hardware question. The O/S does not touch the ram
directly.
> Configurations are:
> SMP Intel 833's running at 133mhz fsb
Drop the speed a bit. You need to find a stable configuration, then
vary one parameter at a time.
> Reg ECC 128's or 256's (Micron chips on 3rd party silicon)
> IBM 9GB, 18GB, and 36GB SCA Ultra 160's
> Kernel's have been compiled with variations of the following:
> Monolithic and Modulated.
> Kernels 2.4.2, 2.4.3, 2.4.3-ac4, 2.4.4-pre1
> SMP
> MTRR support on and off
> 4GB memory support
You don't want that.
> SererWorks IDE Chipset
IDE turned OFF in the bios, no?
> USB on and off
If you can turn it off, turn it off. Also , no IRQ for neither
VGA nor USB, please.
> APM on and off
OFF is correct.
> AIC7XXX drivers (old and new including the latest 6.1.11)
Could be that. You'll have to take the board out to find out!
> With and without Luns
> 5000 ms Initial bus reset delay
> Netfilter on and off
> QOS on and off
> Compilers used:
> kgcc-1.1.2-40
That's right.
> gcc-2.96-69
That's wrong. Don't use it.
Peter
------------------------------
From: "Rainer Traut" <[EMAIL PROTECTED]>
Subject: scsi module aha1542.o doesn't load
Date: Fri, 13 Apr 2001 17:50:27 +0200
Hi *,
system is rh7,
my modules.conf looks like this:
alias scsi_hostadapter aha1542
alias eth0 de4x5
alias eth1 3c90x
# alias scsi_hostadapter aha1542
# alias parport_lowlevel parport_pc
But it seems, like it doesn't even try to load that module
at startup....
An insmod aha1542 works great...
The network modules load fine.
What am I doing wrong?
What script tries to load the scsi modules?
MfG
Rainer
------------------------------
Date: Fri, 13 Apr 2001 11:15:36 -0500
Crossposted-To: comp.os.linux.misc
Subject: Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!!
From: Anonymous <[EMAIL PROTECTED]>
Thanks for responding Peter. I have some notes on your responses below.
"Peter T. Breuer" wrote:
>
> In comp.os.linux.misc Anonymous <[EMAIL PROTECTED]> wrote:
> > We have several new Supermicro 6010 L\H series servers. They work fine
> > until you add more than 2 DIMMs ram. Then for the life of us we can not
> > keep them running. See the set up information below. They are more
>
> > Once installed they will not compile anything without seg faults and
> > core dumps. They kernel panics about every 4 to 8 hours and we have
>
> Set the bios to very conservative values .. and check that the mobo
> accepts the type of ram you have.
>
> > I have made sure that APM is on in the BIOS and CAS is set to 3. I have
>
> ON! You mean OFF, don't you? You can't have APM in an SMP machine.
>
It gets worse if it is off. I only detect 64MB of the 512MB to 1024MB.
> > Is there anything else I might be missing? Distro, Ram, BIOS???
>
> It's purely a hardware question. The O/S does not touch the ram
> directly.
>
> > Configurations are:
> > SMP Intel 833's running at 133mhz fsb
>
> Drop the speed a bit. You need to find a stable configuration, then
> vary one parameter at a time.
>
> > Reg ECC 128's or 256's (Micron chips on 3rd party silicon)
> > IBM 9GB, 18GB, and 36GB SCA Ultra 160's
>
> > Kernel's have been compiled with variations of the following:
> > Monolithic and Modulated.
> > Kernels 2.4.2, 2.4.3, 2.4.3-ac4, 2.4.4-pre1
> > SMP
> > MTRR support on and off
> > 4GB memory support
>
> You don't want that.
On the 1024MB machines I get the following warning in the dmesg:
Can only detect 824MB of Ram. Try recompiling with 4GB support.
>
> > SererWorks IDE Chipset
>
> IDE turned OFF in the bios, no?
Not for the CD-Rom drives.
>
> > USB on and off
>
> If you can turn it off, turn it off. Also , no IRQ for neither
> VGA nor USB, please.
>
> > APM on and off
>
> OFF is correct.
>
> > AIC7XXX drivers (old and new including the latest 6.1.11)
>
> Could be that. You'll have to take the board out to find out!
>
These are on-board chipsets.
> > With and without Luns
> > 5000 ms Initial bus reset delay
> > Netfilter on and off
> > QOS on and off
>
> > Compilers used:
> > kgcc-1.1.2-40
>
> That's right.
>
> > gcc-2.96-69
>
> That's wrong. Don't use it.
>
> Peter
--
************************************
Scott England-Sullivan
Dir. MIS
InterHost USA, Inc.
901 A South Neil Street
Champaign, IL 61820
W. 217-403-1120
E. [EMAIL PROTECTED]
************************************
--------== Posted Anonymously via Newsfeeds.Com ==-------
Featuring the worlds only Anonymous Usenet Server
-----------== http://www.newsfeeds.com ==----------
------------------------------
From: "User Rdkeys Robert D. Keys" <[EMAIL PROTECTED]>
Subject: Any thoughts on a good Sparc Linux?
Date: 13 Apr 2001 16:02:34 GMT
I am thinking of setting up a couple Linux machines, one a server, and
one a desktop, using sun4c sparcs. I see in the archives that there
are sparc distros by SUSE, Debian, Redhat, and Slackware. The last
Linux I used was the old 0.96 stuff from a long time ago. My background
is fairly strong in the BSD side of the family on sparc hardware. But,
I wanted to give Linux a fair shake on this architecture. Can anyone
suggest a few pros and cons of any of the above distros on sparc hardware?
Slackware and SUSE look pretty good. Redhat I have heard mixed things
about. Debian requires a floppy install, initially, apparently, since
I did not find any bootable cd images about, although there seemed to be
some overseas. Is there a url to an official Debian iso image somewhere?
The desktop will be gui driven, and the server tty driven.
Any thoughts or suggestions as to what works relatively well on sun4c
hadware are appreciated.
Thanks
Bob
------------------------------
From: "Troy Bowman" <[EMAIL PROTECTED]>
Crossposted-To: comp.arch.storage,comp.sys.ibm.pc.hardware.storage
Subject: Re: today's harddrives will surely fail before dialup users manage to fill
them up?
Date: Fri, 13 Apr 2001 11:31:26 -0500
Reply-To: "Troy Bowman" <[EMAIL PROTECTED]>
You have not seen my hard drive. I just bought a 45GB drive and have
already filled over 1/2 of it.
"Dan Jacobson" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> It seems that at least for a home dialup user, today's PC hard disks
> would surely fail before he could manage to fill it all up via say,
> browsing, even with caching proxies, etc. Therefore it seems deleting
> unneeded files might become a routine of the past? Hmm, 30 GB /(20
> MB/day)=4 years
> --
> http://www.geocities.com/jidanni Tel886-4-25854780 e-mail:restore .com.
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: Supermicro 6010 + (3to4 DIMMS) + RH 7 + 2.4 = HELP!!
Crossposted-To: comp.os.linux.misc
Date: Fri, 13 Apr 2001 16:39:26 GMT
In comp.os.linux.hardware Anonymous <[EMAIL PROTECTED]> wrote:
> Thanks for responding Peter. I have some notes on your responses below.
> "Peter T. Breuer" wrote:
>>
>> In comp.os.linux.misc Anonymous <[EMAIL PROTECTED]> wrote:
>> > We have several new Supermicro 6010 L\H series servers. They work fine
>> > until you add more than 2 DIMMs ram. Then for the life of us we can not
>> > keep them running. See the set up information below. They are more
>>
>> > Once installed they will not compile anything without seg faults and
>> > core dumps. They kernel panics about every 4 to 8 hours and we have
>>
>> Set the bios to very conservative values .. and check that the mobo
>> accepts the type of ram you have.
>>
>> > I have made sure that APM is on in the BIOS and CAS is set to 3. I have
>>
>> ON! You mean OFF, don't you? You can't have APM in an SMP machine.
> It gets worse if it is off. I only detect 64MB of the 512MB to 1024MB.
That merely tells you that the bios is not compliant, and that
you should not attempt to use it at all, still less for "memory detection"!
Define the amount of memory you want to be seen in the boot parameters
of the kernel, but leave APM off. It is not compatible with SMP.
>> > 4GB memory support
>>
>> You don't want that.
> On the 1024MB machines I get the following warning in the dmesg:
> Can only detect 824MB of Ram. Try recompiling with 4GB support.
Certainly the normal memory split allows you to put in up to about
950MB (1GB minus kernel minus indirection page tables). You shouldn't
need to change that unless you have about 1GB, and you shouldn't be
putting in 1GB just to do basic testing with! get to the basic stable
configuration, and then change things one at a time ...
> Not for the CD-Rom drives.
Take 'em out and turn off IDE: You have t get to a stable configuration
before you can begin testing variations. If that means just ram plus
cpu plus mobo, so be it.
>>
>> OFF is correct.
>>
>> > AIC7XXX drivers (old and new including the latest 6.1.11)
>>
>> Could be that. You'll have to take the board out to find out!
> These are on-board chipsets.
Disable it. Onboard stuff is always flakey.
You must get to a stable configuration. It's not use saying of a heap
of devices that they don't work together! It doesn't help you diagnose
anything. You have to start with zero, and add one thing at a time
until you find the ONE addition that causes the instability.
Developing a surefire test for the instability would also be useful.
When I have had pci instabilities in the past, I have alsways found
that an ssh over the net would kill the machine, for example.
Memory problems in themselves tend to be shown up by kernel
compilations. CPU and DMA problems also are evoked by compilations.
Peter
------------------------------
From: Joshua Baker-LePain <[EMAIL PROTECTED]>
Subject: Re: Random tape device...
Date: 13 Apr 2001 16:57:13 GMT
John Ouellette <[EMAIL PROTECTED]> wrote:
> How stable is devfs? I haven't thought about it in a while, but last
> I remember there were 'issues'.... I'll be upgrading to 2.4.x in
> the near future, so this might be an option.
I can't really speak to the stability. I mucked about with it for a bit
when I installed the SGI XFS pre-release, which had it enabled. This
was RH7, and I had to change some bits in /etc/security/console.perms
in order, e.g., to be able to 'startx'. Everything "seemed" to work after
that (although I didn't try all that much stuff). But I didn't want to fully
explore it (I was more interested in the XFS bits, obviously), and the box
was eventually going into production, so when I compiled a more recent
2.4.3-XFS kernel, I disabled it. There's some discussion of devfs on
the linux-xfs mailing list over at SGI, if you want to look into it more.
BTW, in general, the recommendation seems to be to hold off a bit on
2.4.x unless it's got features you really need. When the 2.5 tree gets
started, that's when the 2.4 tree has been deemed bug-free enough
to not need everybody's attention. That being said, RH7.1 is supposedly
right around the corner, and based on 2.4.
--
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
******************************
