Problems with 2.4.13, kt133a, and IBM DLTA drives

Problems with 2.4.13, kt133a, and IBM DLTA drives

Post by Michael Perr » Thu, 20 Dec 2001 03:50:43



HI all-

I have a system with the VIA kt133 southbridge motherboard/controller
combo and an IBM DLTA drive that is 45g.  The system is a AMD duron
800mhz which does only webserving for me.  If I switch on dma and use
amanda to backup the system, I get a whole hosts of dma restart errors.
However, I can use a Maxtor 40g drive with no problems with the same
system.  The system is running debian testing with a 2.4.13 kernel.

In regular operation, the system is just fine.  I have a system almost
the same which is a AMD 1g with twin maxtor drives which exhibits none
of these problems.

BTW, I am using the VIA workaround ide driver in the kernel.  The only
time I see DMA resets is with amanda running and do a full "dump" or
backup to my tapeserver.  It seems like the dma resets happen under some
degree of system load but only with the IBM drive.  I have extra maxtor
drives which I could use and just migrate the entire system over; but I
would like to understand the how's and why's of this.

If I switch off pci dma in the kernel, I get dismal transfers when
running hdparm -Tt /dev/hda.  This system only serves up webpages so its
not used in some intensive operations; but I had read good things
overall about the IBM drives and found a sale on them a bit ago.

Thanks.

--
Michael Perry | "Do or do not; there is no try" Master Yoda

 
 
 

Problems with 2.4.13, kt133a, and IBM DLTA drives

Post by lobotom » Thu, 20 Dec 2001 04:18:17


The IBM DTLA (75GXP) drives actually have some pretty horrendous
mechanical problems (maybe you've heard?) but if you aren't noticing any
loud "click-click-click" type noises your drive is probably ok.

Its more likely related to the VIA IDE controller.  I've noticed some
problems like this with older drives (with a 693A/596 board, admittedly
older, but uses the same driver) that seem to occur pretty randomly (some
drives work, some don't).  You just have to turn DMA off.  You can
probably get much of the performance benefit by turning on the other
major optimizations with hdparm (32-bit io, multi mode, etc.).


Quote:> HI all-

> I have a system with the VIA kt133 southbridge motherboard/controller
> combo and an IBM DLTA drive that is 45g.  The system is a AMD duron
> 800mhz which does only webserving for me.  If I switch on dma and use
> amanda to backup the system, I get a whole hosts of dma restart errors.
> However, I can use a Maxtor 40g drive with no problems with the same
> system.  The system is running debian testing with a 2.4.13 kernel.

> In regular operation, the system is just fine.  I have a system almost
> the same which is a AMD 1g with twin maxtor drives which exhibits none
> of these problems.

> BTW, I am using the VIA workaround ide driver in the kernel.  The only
> time I see DMA resets is with amanda running and do a full "dump" or
> backup to my tapeserver.  It seems like the dma resets happen under some
> degree of system load but only with the IBM drive.  I have extra maxtor
> drives which I could use and just migrate the entire system over; but I
> would like to understand the how's and why's of this.

> If I switch off pci dma in the kernel, I get dismal transfers when
> running hdparm -Tt /dev/hda.  This system only serves up webpages so its
> not used in some intensive operations; but I had read good things
> overall about the IBM drives and found a sale on them a bit ago.

> Thanks.

--
Not to have been a dupe, that will have been my best possesion, my best
deed, to have been a dupe, wishing I wasn't, thinking I wasn't, knowing
I was, not being a dupe of not being a dupe.  
--Samuel Beckett, The Unnamable

 
 
 

Problems with 2.4.13, kt133a, and IBM DLTA drives

Post by Mr. Anonymou » Tue, 01 Jan 2002 21:19:09




Quote:> If I switch off pci dma in the kernel, I get dismal transfers when
> running hdparm -Tt /dev/hda.  This system only serves up webpages so its
> not used in some intensive operations; but I had read good things
> overall about the IBM drives and found a sale on them a bit ago.

IBM's DLTA drives are junk.  Do yourself a favor and remove it from your
system.  I have 50 of these drives and 30 of them have already died.  It's
not a question of "IF" these drives will fail, it's a question of when it
will fail: today or next week, but it will fail and it will fail very soon.  

For those who don't know IBM is being sued over these drives.  IBM made two
different drives under the same model number: a 75GB and a 76GB.  The 75GB
drives (the first ones manufactured) are good drives, but about March of
2001 IBM changed something and started making 76GB drives.  These drives are  
junk.  I wouldn't use them if someone gave me 100 of them for free.  To make
things worse they normally don't just totally fail, they die slowly to allow
time to corrupt your data.  Badblocks/timeouts just start creaping in.  

Steve.

 
 
 

Problems with 2.4.13, kt133a, and IBM DLTA drives

Post by Brian Brunne » Wed, 02 Jan 2002 07:00:39





>> > If I switch off pci dma in the kernel, I get dismal transfers when
>> > running hdparm -Tt /dev/hda.  This system only serves up webpages so its
>> > not used in some intensive operations; but I had read good things
>> > overall about the IBM drives and found a sale on them a bit ago.

>> IBM's DLTA drives are junk.  

That's DTLA.

Quote:>> Do yourself a favor and remove it from your system.  

Send them to me!
I use two DTLA 307045's in a RAID-0 array FLAWLESSLY for over 18 months.

Quote:>> I have 50 of these drives and 30 of them have already died.  

Nice hammer you got there!

Quote:>> It's not a question of "IF" these drives will fail, it's a question of when it
>> will fail: today or next week, but it will fail and it will fail very soon.

18 months and counting!

Quote:>> For those who don't know IBM is being sued over these drives.  

'Being sued' only means somebody hired a lawyer.

Quote:>> IBM made two different drives under the same model number: a 75GB and a 76GB.  

Hooboy!  Get some facts please! you're replacing the ignorant with the
misinformed, and that's not progress!

The 3070 series came in 15, 30, 45, 60, & 75 GB sizes. Maybe others also.

The 3060 series (which came *after* the 3070 series) comes in different
sizes, I haven't bought any of them (yet).

The rest of this 'information' I dis as disinformation.

Quote:>> The 75GB drives (the first ones manufactured) are good drives, but about March of
>> 2001 IBM changed something and started making 76GB drives.  These drives are  
>> junk.  I wouldn't use them if someone gave me 100 of them for free.  To make
>> things worse they normally don't just totally fail, they die slowly to allow
>> time to corrupt your data.  Badblocks/timeouts just start creaping in.  

>> Steve.

--
Brian Brunner
ABIT KT7-RAID: (http://go.to/kt7faq)

3x128MB Kingmax PC133 CAS2 all mem optomizations (TURBO)
2x307045(RAID-0) (IDE3 & 4 Masters, CS)
Plextor 12x10x32, Zip-100.
Sound Blaster Live Platinum (PCI 2)
US Robotics external V.92 modem.
Win98SE. No Problems.
 
 
 

Problems with 2.4.13, kt133a, and IBM DLTA drives

Post by Michael Perr » Wed, 02 Jan 2002 10:02:39






>>> > If I switch off pci dma in the kernel, I get dismal transfers when
>>> > running hdparm -Tt /dev/hda.  This system only serves up webpages so its
>>> > not used in some intensive operations; but I had read good things
>>> > overall about the IBM drives and found a sale on them a bit ago.

>>> IBM's DLTA drives are junk.  

> That's DTLA.

>>> Do yourself a favor and remove it from your system.  

> Send them to me!
> I use two DTLA 307045's in a RAID-0 array FLAWLESSLY for over 18 months.

>>> I have 50 of these drives and 30 of them have already died.  

> Nice hammer you got there!

>>> It's not a question of "IF" these drives will fail, it's a question of when it
>>> will fail: today or next week, but it will fail and it will fail very soon.

> 18 months and counting!

>>> For those who don't know IBM is being sued over these drives.  

> 'Being sued' only means somebody hired a lawyer.

>>> IBM made two different drives under the same model number: a 75GB and a 76GB.  

> Hooboy!  Get some facts please! you're replacing the ignorant with the
> misinformed, and that's not progress!

> The 3070 series came in 15, 30, 45, 60, & 75 GB sizes. Maybe others also.

> The 3060 series (which came *after* the 3070 series) comes in different
> sizes, I haven't bought any of them (yet).

> The rest of this 'information' I dis as disinformation.

>>> The 75GB drives (the first ones manufactured) are good drives, but about March of
>>> 2001 IBM changed something and started making 76GB drives.  These drives are  
>>> junk.  I wouldn't use them if someone gave me 100 of them for free.  To make
>>> things worse they normally don't just totally fail, they die slowly to allow
>>> time to corrupt your data.  Badblocks/timeouts just start creaping in.  

>>> Steve.

Hi-

Thanks for the information.  The IBM drive I was using I perhaps
misidentified at the getgo unfortunately.  Its a 45g drive and perhaps I
stated the drive model number incorrectly.  Fact remains that this drive
is junk.  It could not do dma whatsoever and when I turned dma off it
couldn't handle that either.  I mean junk in the literal sense.  As I
upgraded the drive to a Maxtor 80g, the drive started clanking and it
would not copy files correctly.  One file cycled a IO error continuously
and it was only a small text file.  

For me, as a reference point, the exact operation that caused the dma
errors no longer happens now on this system.  With a Maxtor 80g drive in
the webserver, the dma errors are gone and I get much better hdparm
results.  What exactly I was doing was using amanda to backup using
"dump" the entire hard disk partitions.  I do this on two other systems
running debian with maxtor drives in them and a bsd box running some
hard disk drive.  I have never rebooted that box since I built it since
it does service as a firewall.

I have two of these drives and if you would the 45g size contributed, I
would definitely be interested :)  I could not give you the "clanker"
though.  I tossed that one.

I replaced the drives with Maxtor 80g Diamondmax drives which have never
given me any problems.  Now amanda backs the drive up correctly and dma
works perfectly.

--
Michael Perry | "Do or do not; there is no try" Master Yoda

 
 
 

1. 2.4.13-ac8: crash on IBM Thinkpad 600x

Dear all,

2.4.13-ac8 unfortunately crashes on my IBM Thinkpad 600x every now and
then by a kernel panic.  Unfortunately, not syncing.

Does anyone has similar experienes and managed a workaround?

--lpr
PS: 2.4.12-ac5 runs perfectly stable ;-)

--
Lukas Ruf                        Swiss Federal Institute of Technology
Office: ETZ-G61.2                             Computer Engineering and
Phone: +41/1/632 7312                        Networks Laboratory (TIK)
Fax:   +41/1/632 1035                                      ETH Zentrum
PGP 2.6: ID D20BA2ED;                                    Gloriastr. 35
Fingerprint 6323 B9BC 9C8E 6563  B477 BADD FEA6 E6B7    CH-8092 Zurich
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in

More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

2. hanging telnets and ftps

3. Tulip driver problem with 2.4.13-ac8

4. First install - LILO prompt doesn't appear

5. Google's mm problem - not reproduced on 2.4.13

6. SCP without a password

7. network card problem in RH 7.2 with kernel 2.4.13

8. Runscript problems

9. Problems with 2.4.13 - modules

10. Upgrading kernel 2.4.9 -> 2.4.13 causes ex2fsck-problems.

11. Memory accounting problem in 2.4.13, 2.4.14pre, and possibly 2.4.14

12. Sound problem: initial squealing: solo1: kernel 2.4.13

13. Kernel 2.4.13 compiling problems..