Linux-Hardware Digest #912, Volume #14 Sun, 17 Jun 01 14:13:12 EDT
Contents:
Best IDE Raid 5? ("Jean-Sebastien Morisset")
Re: Linux, DAT and DOS Tapes (Rod Smith)
Can I use my USB mouse whit suse6.2? ("Tozzy")
Re: Question: Erase partition tables from hard drive (Jakob Kosowski)
Re: Weird IDE Problem with Compaq Deskpro 4000 & RH 6.x (Cowtown)
Mandrake 8.0 freezing with DMA problems, please help! (Yu Di)
Re: Question: Erase partition tables from hard drive (Rod)
Re: How to install a UMAX SCSI clone onto the NCR5380 driver (Colin Brough)
Re: How to install a UMAX SCSI clone onto the NCR5380 driver (Phlip)
Multi user machine (Mojo B. Nichols)
Re: Best IDE Raid 5? (Joshua Baker-LePain)
Re: Linux, DAT and DOS Tapes (Krzysztof Gozdziewski)
Re: Capacity of Dell Server running as a web server? ("Heikki Tuuri")
----------------------------------------------------------------------------
From: "Jean-Sebastien Morisset" <[EMAIL PROTECTED]>
Subject: Best IDE Raid 5?
Date: Sun, 17 Jun 2001 10:24:53 -0400
Can anyone recommend the best (caching, bandwidth, ease of config, etc.) IDE
RAID 5 card for use with Linux 2.2.x?
Thanks,
js.
--
Jean-Sebastien Morisset, Sr. UNIX Admin <[EMAIL PROTECTED]>
Personal Homepage <http://jsmoriss.mvlan.net/>; UNIX, Internet,
Homebrewing, Cigars, PCS, PalmOS, CP2020, and other Fun Stuff...
This is Linux Country. On a quiet night you can hear Windows NT reboot!
------------------------------
From: [EMAIL PROTECTED] (Rod Smith)
Crossposted-To: comp.os.linux.misc
Subject: Re: Linux, DAT and DOS Tapes
Date: Sun, 17 Jun 2001 14:35:18 -0000
Reply-To: [EMAIL PROTECTED]
In article <[EMAIL PROTECTED]>,
Juergen Pfann <[EMAIL PROTECTED]> writes:
>
> b) Using tar from UNIX(like) to UNIX(like) systems, it *is* possible to
> use remote tape drives via /etc/rmt. The syntax is like (e.g. backup):
> "tar cvf remote_machine:/dev/tape_device /" (substitute the hostname
> and the appropriate device name on the remote tape "server".
> Thus, this could be possible with a tar port supporting this.
> Now, I do know at least 3 tar ports for the various MS Win systems.
> 2 of these support _local_ SCSI tapes, no remote support IIRC.
> Another one was built assuming the tape drive is on a remote *NIX
> system - BUT that was a stone-age old 16-bit app designed for Win
> 3.x,
> with a very simple GUI and just supporting _8+3_ filenames :-((.
> The name was some like "WinTar"; I think I saw it on some old
> CICA mirror CD.
> Let's hope someone resumed that work as a Win32 app...
It's also possible to back up a Windows system using a tape drive on a
Linux system in several other ways. There are some commercial network
tape backup packages that explicitly support this. I believe ARKEIA
(http://www.arkeia.com) is one, but I'm not positive of that. Another
option would be to use smbtar, which is part of the Samba package, to
do the job from Linux (you'd need to tell Windows to export a drive),
but if this were done for a WinNT/2K system, you'd lose most of the
file security features of NTFS. (You can accomplish much the same thing
with a standard smbmount or mount of a Windows file share, using the
ordinary tar or other Linux backup software.) If the system with the
tape drive has a big enough hard disk, you could create a tape backup
Samba share, which uses a postexec script to copy files from the share
to a tape drive. You could then use any tape backup software on Windows
that can save data to a file, copy it to the tape backup share, and
Linux/Samba could copy that to a tape. This could be awkward to
restore, but it would work. I'm sure there are other products and
creative solutions.
--
Rod Smith, [EMAIL PROTECTED]
http://www.rodsbooks.com
Author of books on Linux & multi-OS configuration
------------------------------
From: "Tozzy" <[EMAIL PROTECTED]>
Subject: Can I use my USB mouse whit suse6.2?
Date: Sun, 17 Jun 2001 16:37:10 +0200
I'm running suse6.2 and I would like to use my USB Microsoft IteliExplorer
Mouse.
Does any body have any suggestions?
Thanx
------------------------------
From: Jakob Kosowski <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,alt.os.linux.mandrake,comp.sys.ibm.pc.hardware.storage
Subject: Re: Question: Erase partition tables from hard drive
Date: Sun, 17 Jun 2001 16:44:30 +0200
Fenster wrote:
> Deleting partition tables and master boot record from hard drive.
>
> Maybe I am phrasing this incorrectly but I want to get rid of all
> partition tables and boot records on a 10 GIG Western Digital hard drive.
> I created a mess when I installed Linux and Windows98 on this drive and
> would like to get it back to the way it was when I first bought it......or
> close to it!
>
> I have used Partition Magic 5.0 and it doesn't help much.
> Is there a Linux program that takes care of such problems?
>
> Any help is appreciated. Posted here or sent to my e-mail.
>
> Thanks
> Fenster
>
>
>
dd if=/dev/zero of=/dev/hda bs=512 count=1
Be careful.
--
"One World, One Web, One Program" - Microsoft Promotional Ad
"Ein Volk, Ein Reich, Ein Führer" - Adolf Hitler
------------------------------
From: [EMAIL PROTECTED] (Cowtown)
Crossposted-To: linux.redhat
Subject: Re: Weird IDE Problem with Compaq Deskpro 4000 & RH 6.x
Date: Sun, 17 Jun 2001 14:31:54 GMT
You missed it Doug. The partition does not hold inof for the bios but
i part of a restore partition that also hold various compaq utilities.
The partition is not needed and can be removed. You really wont need
to utilities with linux. fdisk the whole drive an try again.
On Sat, 16 Jun 2001 15:32:19 GMT, "Doug Holtz"
<[EMAIL PROTECTED]> wrote:
>John;
>Compaq puts a non-dos partition on the front of their drives that holds
>configurations. (It's not on the bios chip). Compaq did not use Fuji
>drives; they bought drives and put their name on it. There fore I suspect
>this is an aftermarket drive.
>Get the Compaq utility that reinstalls the partition for writing config info
>to and put that on and see if it helps the install.
>Doug
>
>John Roy <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> Hi All,
>> I've tried for days to get information on this problem but seems I'm the
>> only persons got it - seems other
>> people with my hardware don't have a problem.
>>
>> I got a cheap used Compaq Despro 4000 (with P166MMX chip) - I've tried
>> to install both RH6.1 and RH6.0 - something I've done many times but
>> here I'm stumped. I put the CD in local CDROM, boot and from there try a
>> typical but minimal install. After the HD has been partitioned and
>> during writting the packages to disk the system just hangs someplace
>> between 0% complete and 95% complete - the other 25% of the time I get a
>> kernel panic with dump of registers (killing interrupt handler - attempt
>> to kill the idle task...).
>>
>> I suspected problems with the HD and do see linux reporting different
>> C/H/S than bios setup - typically bios says 777/128/63 (for 3209MG disk,
>> translation on), whereas linux on bootup reports 847/128/63. I've tried
>> everything - three seperate harddisks (3GB, 4GB and 830MB), LBA
>> translation on and off, any number of combinations of using the
>> hda=C,H,S boot option to force the C/H/S under the drive capacity. All
>> with the
>> same result. DOS seems to install just fine. Oh, I've also upgraded the
>> flash BIOS from compaq and tried every last one of thier utilities.
>>
>> The only clues I have are 1. When I use the hda=C,H,S option and reduce
>> the capacity of the driver I'm typically able to get a lot further, but
>> never more than 95% installed - and that was with a 827/32/63 drive
>> set to 400/15/63. Also ONCE the system hung while reading the CDROM
>> before doing anything much with
>> the hda (smells like the IDE controller right?).
>>
>> Please help I'm stumped - as I said this is a used system from 1997 - so
>> I don't have all info -
>> P166MMX processor - builtin in network card. 1 HD only on primary IDE
>> interface - 3209MB Fuji...
>> (though I've tried 2 other disks), 1 CDROM drive only on the secondary
>> IDE. 32MB Ram, IDE
>> controller and Ethnernet on mainboard.
>>
>> I only want to use this machine as a firewall - so if I can get around
>> this HD problem I'll be happy -
>>
>> Sorry its difficult to give dumps of syslog, but I've examined carefully
>> and they seem no different from
>> other machines I have - except for anything noted above - any ideas?,
>> any magic boot options?, think
>> my IDE controller or motherboard is fried - (I even tried replacing the
>> IDE rippon).
>>
>> Thanks,
>> John
>>
>>
>
------------------------------
From: [EMAIL PROTECTED] (Yu Di)
Crossposted-To: comp.os.linux.setup,alt.os.linux.mandrake
Subject: Mandrake 8.0 freezing with DMA problems, please help!
Date: 17 Jun 2001 08:17:57 -0700
Hi, I have installed Mandrake 8.0 with Cheapbytes CD-Roms. I have
experienced innumerous system freezes during the installation and
after I got it installed. Usually the pattern is that during some RPM
software installation, the system suddenly freezes and I have to press
"reset".
This morning, I was trying to install a package using "rpm --install
..." in the console mode. I have already successfully installed some
other packages in the same session. When the installation starts,
after a while, I got these messages:
hda: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hda: irq timeout: Status=0x58 {DriveReady SeekComplete
DataRequest}
hdb: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hdb: irq timeout: Status=0x50 {DriveReady SeekComplete}
Then the messages repeat themselves until I used "kill -9" in another
console (the system is not yet completely frozen), and I got:
hda: DMA disabled
hdb: DMA disabled
ide0: reset: success
I then ran the rpm command again, and immediately got a system freeze,
the message is:
hda: lost interrupt
So I think this is a DMA problem, am I right? How shall I make it
work? Thanks!
My hard drive information is:
hda is JV100-31 Samsung SV0432A, Ultra DMA Mode-2, S.M.A.R.T. Capable
and Status Ok
hdb is BAC51KJ0 Maxtor 32049H3, Ultra DMA Mode-4, S.M.A.R.T. Capable
and Status Ok
hda: Size=4312, Cyln=8912, Head=15, Wpcom=0, Sec=63, LBA=On, Blk=On,
PIO=4, 32bit=Off
hdb: Size=20492, Cyln=39704, Head=16, Wpcom=0, Sec=63, LBA=On, Blk=On,
PIO=4, 32bit=Off
My processor is PII350. I have 288M memory (8M of which is used as
video card memory)
Thanks a lot!
------------------------------
From: Rod <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,alt.os.linux.mandrake,comp.sys.ibm.pc.hardware.storage
Subject: Re: Question: Erase partition tables from hard drive
Date: Sun, 17 Jun 2001 15:27:15 GMT
Lars Luthman wrote:
> Fenster wrote:
> > Maybe I am phrasing this incorrectly but I want to get rid of all partition
> > tables and boot records on a 10 GIG Western Digital hard drive. I created a
> > mess when I installed Linux and Windows98 on this drive and would like to
> > get it back to the way it was when I first bought it......or close to it!
>
> How did it look when you bought it? If you want to install Linux on it,
> you could simply boot from the install floppy and run the installation
> program (if you have one). I know that Mandrake's installer includes a
> tool that allows you to repartition the hard drive. I think you can use
> it to create FAT partitions too, if you want to install Windows
> afterwards.
>
> --ll
You should install Windows before linux if you plan to have both installed
shouldn't you??
------------------------------
From: [EMAIL PROTECTED] (Colin Brough)
Subject: Re: How to install a UMAX SCSI clone onto the NCR5380 driver
Date: Sun, 17 Jun 2001 16:35:15 +0100
Reply-To: [EMAIL PROTECTED]
On 17 Jun 2001 01:52:09 GMT, Phlip wrote:
> Just to stash the info somewhere...
>
> The 970160-08 ( per http://support.umax.co.uk/technotes/f096B.htm ) has its
> JR1 soldiered shut. This means it does not do IRQ. Its io address should be
> 0x220 or 0x340. The following two lines don't whack my box:
>
> modprobe g_NCR5380 ncr_addr=0x220 dtc_3181e=1
> modprobe g_NCR5380 ncr_addr=0x340 dtc_3181e=1
>
> They also don't return anything, so now I'm off to 'man modprobe' to see
> what a nil output means here...
(I missed the start of this thread, so I'm not absolutely sure of the
context; hope this is relevant.)
I have successfully[1] used the SCSI card that came with my Umax 610s
scanner. Umax call the card I have a UDS-IS-11 SCSI controller card,
which is in fact an OEM version of Domex/DTC Technology's DTC-3181
SCSI card. This is supported by the 'g_NCR5380.c' driver in the latest
kernel versions (certainly from 2.2.0-pre1, probably earlier than that
too). Mustek ship the same card with some of their scanners, and it
was a Mustek owner who patched the generic NCR driver to get it
working.
To use (assuming you've followed the normal instructions for setting
up SCSI!): add the following to your kernel command line:
dtc3181e=0x280,254
(No, I don't remember what the 254 does! Sorry.) When I used this I
compiled it into the kernel, but that was back around the time of
2.2.0; modules stuff has changed since then.
[1] Performance sucks - the card has no interrupts, so during scans
performance is terrible. I got occasional lock ups, I presume due to
the driver, but it was liveable with if scanning is not your main line
of business! In that case get another SCSI card for its better
performance alone! (I now have an Adaptec card.)
--
Cheers
Colin
ps Remove 'REMOVE.' from Reply-To field to reply.
======================================================================
Colin Brough [EMAIL PROTECTED]
------------------------------
From: Phlip <[EMAIL PROTECTED]>
Subject: Re: How to install a UMAX SCSI clone onto the NCR5380 driver
Date: 17 Jun 2001 15:56:54 GMT
Proclaimed Colin Brough from the mountaintops:
> (I missed the start of this thread, so I'm not absolutely sure of the
> context; hope this is relevant.)
The context is...
UMAX Astra 600S ->
dtc3181e chip ->
UDS-IS11 970160-08 card ->
g_NCR5380 driver ->
??? ->
??? ->
sane -> xsane
> I have successfully[1] used the SCSI card that came with my Umax 610s
> scanner. Umax call the card I have a UDS-IS-11 SCSI controller card,
> which is in fact an OEM version of Domex/DTC Technology's DTC-3181
> SCSI card. This is supported by the 'g_NCR5380.c' driver in the latest
> kernel versions (certainly from 2.2.0-pre1, probably earlier than that
> too). Mustek ship the same card with some of their scanners, and it
> was a Mustek owner who patched the generic NCR driver to get it
> working.
Da. I added...
alias scsi_hostadapter g_NCR5380
options g_NCR5380 ncr_addr=0x340 dtc_3181e=1
...to modules.conf, and the system boots up with that okay.
> To use (assuming you've followed the normal instructions for setting
> up SCSI!):
There are instructions for setting up SCSI?
> add the following to your kernel command line:
>
> dtc3181e=0x280,254
If I did that (or if my tweak to modules.conf worked), how would I know?
Would a new entry appear in /dev? Would an existing entry change its
properties?
Should I run sgcheck?
The first ??? in my list represents I don't know what a live g_NCR5380
"looks like" to the rest of the 'puter.
(If so, how do I find sgcheck? Where is its home page?)
If whatever check I can do works, how to I tell SANE about the scanner? Its
copious documentation seems to start after this point.
The second ??? in my list represents I don't know how to declare the
scanner to SANE.
And how on earth are civilians expected to configure this stuff?
--
Phlip [EMAIL PROTECTED]
============== http://phlip.webjump.com ==============
-- All sensors report Patti having a very good time --
------------------------------
From: [EMAIL PROTECTED] (Mojo B. Nichols)
Subject: Multi user machine
Date: 17 Jun 2001 11:08:31 -0500
Does anybody have any experience with using two monitors and keyboards
on one machine? I would like to have two seperate people logged into
their respective accounts on the same machine. The second machine
would be for my girlfriend. I'm thinking I could sink 1/2 what I would
spend on a new machine into the monitor, and make her much happier
then buying a new machine. It seems like this would be a very common
given the number of people in a typical household. And the number of
cheapskate fathers who are tired of buying more crap for their kids;-)
Thanks in advance,
Mojo Nichols
------------------------------
From: Joshua Baker-LePain <[EMAIL PROTECTED]>
Subject: Re: Best IDE Raid 5?
Date: 17 Jun 2001 17:21:57 GMT
Jean-Sebastien Morisset <[EMAIL PROTECTED]> wrote:
> Can anyone recommend the best (caching, bandwidth, ease of config, etc.) IDE
> RAID 5 card for use with Linux 2.2.x?
The canonical suggestion seems to be the 3ware cards.
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
------------------------------
From: Krzysztof Gozdziewski <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Linux, DAT and DOS Tapes
Date: 17 Jun 2001 18:04:41 GMT
In comp.os.linux.hardware Peter Kohut <[EMAIL PROTECTED]> wrote:
> I have a machine running Redhat Linux with a SCSI interface and a SONY 4mm
> DAT hanging of it. I also have a bunch of tapes which were originally
> written using NT Backup and Restore. My question is
> a) is there any way that I can restore the content of the tapes using Linux?
> b) I also have a second machine running Win 2000, but unfortunately without
> SCSI. The Win 2000 machine is connected to the Linux machine using simple
> TCP/IP and SAMBA shares. Is it possible to configure the systems such that I
> can use the Windows 2000 Backup/Restore application with the DAT drive
> hanging of the Linux machine?
You can use free backup software called amanda
http://www.amanda.org/
Amanda works in serwer/client architecure. Windows machines may be
archived with amanda too, it needs samba on the serwer working.
Take a look at http://www.backupcentral.com/amanda.html.
I think it will perfectly fit your needs.
Regards,
--
Krzysztof Gozdziewski
Life ain't fair, but the root password helps - BOFH
------------------------------
From: "Heikki Tuuri" <[EMAIL PROTECTED]>
Subject: Re: Capacity of Dell Server running as a web server?
Date: Sun, 17 Jun 2001 18:08:47 GMT
Hi!
Steve Wolfe wrote in message ...
>> > No, I didn't say that it was the bottleneck. To pretend that
>anything
>> >that is not a "bottleneck" can be dropped from the equation isn't
>correct.
>> >And trying to equate sheer capacity with web-serving capacity isn't
>> >correct. Even though the database machine may not be anywhere near
>> >capacity, it is still part of the equation.
>>
>>
>> How many DB queries your application typically does per hit?
>
> Depending on the page, it can be anywhere from one to tens. But what I
>was referring to is that the overall latency of the request is dependant
>on several things - of which the database is one, but not the only factor.
Do you know what is the impact of PHP, Perl, or the scripting language
you are using? What does 'top' say about the CPU usage of Apache,
PostgreSQL and other processes?
If you use sockets and write your queries through the C API, then you should
be able to make at least 1000, maybe 10000 simple database queries per
second. That translates to serving 100 to 1000 page hits per second.
The Perl API I used in some tests was very slow: it ate up to 67 % of CPU
capacity. PHP might be even slower.
>> If you can run the tests posted on my website or other tests in an
>> impartial way on PostgreSQL and MySQL/InnoDB, I am very interested
>> to hear the results. I can help in setting InnoDB parameters, though
>> probably the examples in the InnoDB manual at http://www.innodb.com
>> suffice to show how to set them in the optimal way. I have sought to
>find
>> solutions which make tuning easy.
>
> I ran a couple of the tests, and found results that were somewhat worse
>or a lot better - but the point I'm trying to make is that your tests
>aren't really indiciative of real-world applications, in my opinion, and
>the systems haven't necessarily been set up in a "real-world"
>configuration.
I interpret that you mean the result numbers I have posted at my website
are approximately right also when PostgreSQL is tuned optimally?
I am busy writing arbitrary size BLOB support for InnoDB. I have put only
little time to benchmarking, and have been hoping some user would provide
numbers from a real-world workload. The user I quoted earlier talked about
scaling to at least 200 users (threads) and 900 queries per second.
> The other issue is, of course, scaling. Rather than simply giving the #
>of transactions/second at a single # of simultaneous transactions,
>plotting a graph of TPS vs. connections would illustrate the trends in a
>much better manner.
I have studied scaling a lot. I have looked for ways to reduce
semaphore contention inside the server, like I say on the 'Features' page
at http://www.innodb.com. Inside a DB engine scaling is restricted
either by semaphore contention or saturation of the memory bus
in the computer. Outside the DB server, operating system
inefficiency in communication handling and thread scheduling can
hamper performance, but we DB people do not have too much control
over that. My experience is that single thread tests give a good
indication also of multithread performance.
I measured today the following scaling numbers on the 2-CPU Xeon
450 MHz, 512 MB RAM computer running Linux kernel 2.4. I configured
128 MB buffer pool to InnoDB. In the insert test I made 20 copies of
the insert Perl program I have listed on my benchmark web page,
changing the table name in the i'th program to 'speedi'. In the select test
I just create n processes, each of which runs the Perl select test program
on my web page, using the primary key to fetch the rows. The insert
speed is calculated by dividing the total number of inserts done by the
threads by the wall clock time elapsed from the start of the first thread
to the completion of the last client thread.
Run on MySQL/InnoDB-3.23.39:
number of concurrent client threads
1 2 5 10
20 50
--------------------------------------------------------
======================
insert speed
rows / sec 4300 5000 6100 6250 7400
select speed
rows /sec 1750 3030 3330 3330 3290
3150
Thus we see that there are no scalability problems seen with respect
to the number of concurrent client threads. Note that these tests were
CPU-bound, and in the insert test 'top' showed the DB server using 2/3
of CPU time, the Perl processes 1/3. In the select test the DB server
used 1/3 of CPU time and the Perl processes 2/3.
Regards,
Heikki
>steve
>
>
>
------------------------------
** 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
******************************
