Via VT8233C + Western Digital Caviar = corruption hell?

Via VT8233C + Western Digital Caviar = corruption hell?

Post by Michael E. Stora, Ph. » Wed, 22 May 2002 03:44:00



I am having tremendous difficulties with my Linux box based on a Soyo
Dragon Plus motherboard using two 60 GB Western Digital drives.

I must install Linux with the motherboard on failsafe setting, make
several attempts (sometimes dozens), and still go back and repair
random packages that returned errors.  Once installed, file system
corruption is frequent (using Reiser, btw) and copying large/complex
directories between master and slave using cp -a cause segfaults and
or freezes that severely corrupt the file system, so I canno backup my
machine to protect me from all this corruption.  The fact that a crash
will occur when copying /usr is certain, occasionally for /opt or
/home, however the files that cannot stat are always different each
time.

I've tried 4 sets of hard drive cables, many settings (DMA off does
not seem to help, interupt unmasking off seems to prevent lockup but
data/file system corruption still occurs when copying).

I've tried two motherboards, 4 IDE cables, and SuSE 7.3 and 8.0, two
power supplies, etc.  I'm now debating trying different hard drives.
This is getting very expensive!

I am really getting desperate as I have been struggling with this for
months with no useful help from Soyo tech support or Western Digital.

Meanwhile I've tried to "STFW" and come up with a lot of rumors about
problems with Via chip sets and/or Western Digital drives (often
posted by some big names in the Liunx community) often to be followed
up by a denial from someone equally well-known.  I've also seen some
words to the effect that kernel versions lower than 2.5.2 do not
support the VT8233 southbridge, however everything seems to be
recognized on my system.

Has anyone had similar problems or can point me in the right direction
as I try to research this issue?

Mike

 
 
 

Via VT8233C + Western Digital Caviar = corruption hell?

Post by The On » Wed, 22 May 2002 05:08:02


well, i'm experiencing absolutely the same problem (data corruption, lockups
on high load, sometimes i can't even listen to mp3s without a lockup,...)
 i have the same chipset (the 8233 southbridge), but another mainboard (abit
kr7a-raid)

i'm running a ftp-server on that machine but of course not at the moment,
cuz file corruption is simply unbearable. i have 4 hdds (all are western
digital (80-100GB, 7200rpm)....hmmm) attached to the onboard raid-controller
and those 4 have this data corruption. the one disk (seagate barracuda IV
80) where linux is installed is attached to the standard ide-controller. i
don't have probs with this disk fortunately - at least no data corruption
(lockups when listening to mp3s from that disk also appear) - but i don't
have much load there, the other disks have a constant high load cuz of the
ftpserver (internal network) and me working on them.

i have tried many things (different kernels (though not yet a 2.5.x),
patches (mainly from alan cox), cables, and so on...), another pci
raid-controller (highpoint 370), but the problem still persists.
i don't think that any other hardware (PSU, RAM,...) is related to that
problem.

i also have windows installed on the 80gb seagate and the prob doesn't occur
there. but i guess it's cuz of the thousands of VIA patches you need
there...(e.g. pci latency patch)
i think it's a problem with the pci bus and the southbridge, what you could
try is the patches by VIA:
http://www.veryComputer.com/(ide storage drivers)

i didn't try them yet, i'm really [...] off by VIA and i don't have the time
anymore to play with this VIA *
by tomorrow i won't have a VIA board anymore! i've spent too much time and i
can't have my ftp offline all the time. i'm working 90-95% of the time under
linux where i can't even save my data to the other disks without having some
data corruption.
i'll buy a MSI 745 ultra with sis 745 chipset with a rocketraid 404
raid-controller.

i'd be interested if you find any solution and i'm gonna tell ya if the
probs go away with another chipset (will take some time though, cuz the
rocketraid wasn't on stock)
i'm very happy with my ecs k7s5a (sis 735) in one of my other machines so i
hope there are no probs with the sis745 ;)

the one



Quote:> I am having tremendous difficulties with my Linux box based on a Soyo
> Dragon Plus motherboard using two 60 GB Western Digital drives.

> I must install Linux with the motherboard on failsafe setting, make
> several attempts (sometimes dozens), and still go back and repair
> random packages that returned errors.  Once installed, file system
> corruption is frequent (using Reiser, btw) and copying large/complex
> directories between master and slave using cp -a cause segfaults and
> or freezes that severely corrupt the file system, so I canno backup my
> machine to protect me from all this corruption.  The fact that a crash
> will occur when copying /usr is certain, occasionally for /opt or
> /home, however the files that cannot stat are always different each
> time.

> I've tried 4 sets of hard drive cables, many settings (DMA off does
> not seem to help, interupt unmasking off seems to prevent lockup but
> data/file system corruption still occurs when copying).

> I've tried two motherboards, 4 IDE cables, and SuSE 7.3 and 8.0, two
> power supplies, etc.  I'm now debating trying different hard drives.
> This is getting very expensive!

> I am really getting desperate as I have been struggling with this for
> months with no useful help from Soyo tech support or Western Digital.

> Meanwhile I've tried to "STFW" and come up with a lot of rumors about
> problems with Via chip sets and/or Western Digital drives (often
> posted by some big names in the Liunx community) often to be followed
> up by a denial from someone equally well-known.  I've also seen some
> words to the effect that kernel versions lower than 2.5.2 do not
> support the VT8233 southbridge, however everything seems to be
> recognized on my system.

> Has anyone had similar problems or can point me in the right direction
> as I try to research this issue?

> Mike


 
 
 

Via VT8233C + Western Digital Caviar = corruption hell?

Post by Michael E. Stora, Ph. » Thu, 23 May 2002 01:01:00


The only thing I can find that helps dramatically is reducing the
memory clock from 133 to 100.  Have you tried this? Turning DMA off
does nothing, turning interrupt masking off only lessens the change of
a hard lockup (more likely to segfault) but the drive is still
corrupted if I try to write a big chunk to it.

Mike


> well, i'm experiencing absolutely the same problem (data corruption, lockups
> on high load, sometimes i can't even listen to mp3s without a lockup,...)
>  i have the same chipset (the 8233 southbridge), but another mainboard (abit
> kr7a-raid)

> i'm running a ftp-server on that machine but of course not at the moment,
> cuz file corruption is simply unbearable. i have 4 hdds (all are western
> digital (80-100GB, 7200rpm)....hmmm) attached to the onboard raid-controller
> and those 4 have this data corruption. the one disk (seagate barracuda IV
> 80) where linux is installed is attached to the standard ide-controller. i
> don't have probs with this disk fortunately - at least no data corruption
> (lockups when listening to mp3s from that disk also appear) - but i don't
> have much load there, the other disks have a constant high load cuz of the
> ftpserver (internal network) and me working on them.

> i have tried many things (different kernels (though not yet a 2.5.x),
> patches (mainly from alan cox), cables, and so on...), another pci
> raid-controller (highpoint 370), but the problem still persists.
> i don't think that any other hardware (PSU, RAM,...) is related to that
> problem.

> i also have windows installed on the 80gb seagate and the prob doesn't occur
> there. but i guess it's cuz of the thousands of VIA patches you need
> there...(e.g. pci latency patch)
> i think it's a problem with the pci bus and the southbridge, what you could
> try is the patches by VIA:
> http://www.veryComputer.com/(ide storage drivers)

> i didn't try them yet, i'm really [...] off by VIA and i don't have the time
> anymore to play with this VIA *
> by tomorrow i won't have a VIA board anymore! i've spent too much time and i
> can't have my ftp offline all the time. i'm working 90-95% of the time under
> linux where i can't even save my data to the other disks without having some
> data corruption.
> i'll buy a MSI 745 ultra with sis 745 chipset with a rocketraid 404
> raid-controller.

> i'd be interested if you find any solution and i'm gonna tell ya if the
> probs go away with another chipset (will take some time though, cuz the
> rocketraid wasn't on stock)
> i'm very happy with my ecs k7s5a (sis 735) in one of my other machines so i
> hope there are no probs with the sis745 ;)

> the one



> > I am having tremendous difficulties with my Linux box based on a Soyo
> > Dragon Plus motherboard using two 60 GB Western Digital drives.

> > I must install Linux with the motherboard on failsafe setting, make
> > several attempts (sometimes dozens), and still go back and repair
> > random packages that returned errors.  Once installed, file system
> > corruption is frequent (using Reiser, btw) and copying large/complex
> > directories between master and slave using cp -a cause segfaults and
> > or freezes that severely corrupt the file system, so I canno backup my
> > machine to protect me from all this corruption.  The fact that a crash
> > will occur when copying /usr is certain, occasionally for /opt or
> > /home, however the files that cannot stat are always different each
> > time.

> > I've tried 4 sets of hard drive cables, many settings (DMA off does
> > not seem to help, interupt unmasking off seems to prevent lockup but
> > data/file system corruption still occurs when copying).

> > I've tried two motherboards, 4 IDE cables, and SuSE 7.3 and 8.0, two
> > power supplies, etc.  I'm now debating trying different hard drives.
> > This is getting very expensive!

> > I am really getting desperate as I have been struggling with this for
> > months with no useful help from Soyo tech support or Western Digital.

> > Meanwhile I've tried to "STFW" and come up with a lot of rumors about
> > problems with Via chip sets and/or Western Digital drives (often
> > posted by some big names in the Liunx community) often to be followed
> > up by a denial from someone equally well-known.  I've also seen some
> > words to the effect that kernel versions lower than 2.5.2 do not
> > support the VT8233 southbridge, however everything seems to be
> > recognized on my system.

> > Has anyone had similar problems or can point me in the right direction
> > as I try to research this issue?

> > Mike

 
 
 

Via VT8233C + Western Digital Caviar = corruption hell?

Post by Michael E. Stora, Ph. » Thu, 23 May 2002 07:37:21



> well, i'm experiencing absolutely the same problem (data corruption, lockups
> on high load, sometimes i can't even listen to mp3s without a lockup,...)
>  i have the same chipset (the 8233 southbridge), but another mainboard (abit
> kr7a-raid)

> i'm running a ftp-server on that machine but of course not at the moment,
> cuz file corruption is simply unbearable. i have 4 hdds (all are western
> digital (80-100GB, 7200rpm)....hmmm) attached to the onboard raid-controller
> and those 4 have this data corruption. the one disk (seagate barracuda IV
> 80) where linux is installed is attached to the standard ide-controller. i
> don't have probs with this disk fortunately - at least no data corruption
> (lockups when listening to mp3s from that disk also appear) - but i don't
> have much load there, the other disks have a constant high load cuz of the
> ftpserver (internal network) and me working on them.

> i have tried many things (different kernels (though not yet a 2.5.x),
> patches (mainly from alan cox), cables, and so on...), another pci
> raid-controller (highpoint 370), but the problem still persists.
> i don't think that any other hardware (PSU, RAM,...) is related to that
> problem.

> i also have windows installed on the 80gb seagate and the prob doesn't occur
> there. but i guess it's cuz of the thousands of VIA patches you need
> there...(e.g. pci latency patch)
> i think it's a problem with the pci bus and the southbridge, what you could
> try is the patches by VIA:
> http://www.veryComputer.com/(ide storage drivers)

> i didn't try them yet, i'm really [...] off by VIA and i don't have the time
> anymore to play with this VIA *
> by tomorrow i won't have a VIA board anymore! i've spent too much time and i
> can't have my ftp offline all the time. i'm working 90-95% of the time under
> linux where i can't even save my data to the other disks without having some
> data corruption.
> i'll buy a MSI 745 ultra with sis 745 chipset with a rocketraid 404
> raid-controller.

> i'd be interested if you find any solution and i'm gonna tell ya if the
> probs go away with another chipset (will take some time though, cuz the
> rocketraid wasn't on stock)
> i'm very happy with my ecs k7s5a (sis 735) in one of my other machines so i
> hope there are no probs with the sis745 ;)

> the one



> > I am having tremendous difficulties with my Linux box based on a Soyo
> > Dragon Plus motherboard using two 60 GB Western Digital drives.

> > I must install Linux with the motherboard on failsafe setting, make
> > several attempts (sometimes dozens), and still go back and repair
> > random packages that returned errors.  Once installed, file system
> > corruption is frequent (using Reiser, btw) and copying large/complex
> > directories between master and slave using cp -a cause segfaults and
> > or freezes that severely corrupt the file system, so I canno backup my
> > machine to protect me from all this corruption.  The fact that a crash
> > will occur when copying /usr is certain, occasionally for /opt or
> > /home, however the files that cannot stat are always different each
> > time.

> > I've tried 4 sets of hard drive cables, many settings (DMA off does
> > not seem to help, interupt unmasking off seems to prevent lockup but
> > data/file system corruption still occurs when copying).

> > I've tried two motherboards, 4 IDE cables, and SuSE 7.3 and 8.0, two
> > power supplies, etc.  I'm now debating trying different hard drives.
> > This is getting very expensive!

> > I am really getting desperate as I have been struggling with this for
> > months with no useful help from Soyo tech support or Western Digital.

> > Meanwhile I've tried to "STFW" and come up with a lot of rumors about
> > problems with Via chip sets and/or Western Digital drives (often
> > posted by some big names in the Liunx community) often to be followed
> > up by a denial from someone equally well-known.  I've also seen some
> > words to the effect that kernel versions lower than 2.5.2 do not
> > support the VT8233 southbridge, however everything seems to be
> > recognized on my system.

> > Has anyone had similar problems or can point me in the right direction
> > as I try to research this issue?

> > Mike

I'm going to try keeping the mobo since my sound, USB, and LAN work
great. I picked up a RocketRAID 404 from newegg.com for $100.  Never
did get the onboard Promise RAID to work under Linux (it works in
normal IDE mode but is not bootable).  I'll move the hard drives to
the card, see if the CD-RW and DVD work ok on the main IDE (it might
be a VIA/Western Digital interraction), etc.  I have the whole week of
7/4 off from work so I'll have plenty of time to tinker.

Mike

 
 
 

Via VT8233C + Western Digital Caviar = corruption hell?

Post by The On » Sun, 26 May 2002 04:09:15




Quote:> The only thing I can find that helps dramatically is reducing the
> memory clock from 133 to 100.  Have you tried this? Turning DMA off
> does nothing, turning interrupt masking off only lessens the change of
> a hard lockup (more likely to segfault) but the drive is still
> corrupted if I try to write a big chunk to it.

no, i didn't try reducing the memory clock, i also didn't wanna turn DMA
off, this is no solution for me because i'm running a server on that machine
and in PIO mode it would be too slow.

i now have the MSI 745 ultra and i don't have any problems at the moment,
but i still don't have my new rocketraid 404 so i'm currently testing it
with not all drives attached to my old hpt370 (abit hotrod 100).
i have copied many big files to and from the hard disks, uploaded and
downloaded some stuff from the ftp-server, no problems so far.... ;))

i'll post again when i have my rocketraid controller, i dunno how long it
takes...

the one

 
 
 

Via VT8233C + Western Digital Caviar = corruption hell?

Post by Michael E. Stora, Ph. » Sat, 01 Jun 2002 13:02:25





> > The only thing I can find that helps dramatically is reducing the
> > memory clock from 133 to 100.  Have you tried this? Turning DMA off
> > does nothing, turning interrupt masking off only lessens the change of
> > a hard lockup (more likely to segfault) but the drive is still
> > corrupted if I try to write a big chunk to it.

> no, i didn't try reducing the memory clock, i also didn't wanna turn DMA
> off, this is no solution for me because i'm running a server on that machine
> and in PIO mode it would be too slow.

> i now have the MSI 745 ultra and i don't have any problems at the moment,
> but i still don't have my new rocketraid 404 so i'm currently testing it
> with not all drives attached to my old hpt370 (abit hotrod 100).
> i have copied many big files to and from the hard disks, uploaded and
> downloaded some stuff from the ftp-server, no problems so far.... ;))

> i'll post again when i have my rocketraid controller, i dunno how long it
> takes...

> the one

I solved my IDE corruption problems (another post).  In the meanwhile,
my rocketraid 404 arrived.  I compiled the open source drivers on my
laptop, put it on a floppy, booted off the install CD, switched to a
console, loaded the module, then used the installer's "boot installed
system option".  Then I copied  the diver to /lib/modules... rebuilt
my initrd, and ran lilo.  Everything works great and I am getting
reads of almost 38 MB/s from my raid 1 array.

Mike

 
 
 

1. Western Digital Caviar 1.2 GB - Which harddrive to buy?



I have 2 WD 1.08 GB drives--one a 310XX and the other a 210XX (I'm not
certain of the exact model numbers).  Both have 128k buffers.  Both are
very fast, if very slightly noisy, and I haven't had any problems whatever
with either (I've had the 31000 for a year and the 21000 for just a few
weeks).  These are EIDE drives, and I'm running the 31000 on the EIDE
interface and the  21000 on the IDE (secondary) interface of my Premiere II
motherboard (the CDROM is also on the EIDE interface).  I don't know if
this is relevent towards the 1.2 GB drive, but even if it isn't this info
ought to help someone.

From various documents I get the impression that you might have problems
if you use the DOS driver software provided by WD.  Possibly it doesn't
recognize Linux partitions if you have them on the same disk as DOS.
However, I believe that this software is required only if you have an
older computer that requires disk address translation; certainly I'm
not using the WD software and I run DOS and Linux off the same disk.


2. OnStream ADR Tape drivers (DI30) now available

3. Western Digital Caviar HD and 2048 Cylinders in Linux

4. Networking in kickstart file

5. Western Digital Caviar HD and 2484 Cylinders in Linux

6. So slow..Why.

7. Western Digital Caviar 1.2 GB - Which harddrive to buy?

8. Linux Terminal Driver Problem????

9. Western Digital Caviar 1 gig & Busmouse problems :) (sorry)

10. Western Digital 80GB Caviar Hard Drive with 8MB Cache v.s Linux

11. western digital caviar and linux

12. Drive Survey (Western Digital Caviar)

13. western digital caviar and linux