|
Hi,
I'm trying to narrow down the problem some people
have with the ataraid driver for HPT and PROMISE chipsets. If somebody can help
us... hopefully those details will help in far better way.
My setup:
I have a system with an ASUS A7V266-E equipped with
the promise 20265 RAID.
On the VIA KT266A chipset I have:
ide0: DVD-ROM asus E616 ATA100,master,
/dev/hda
ide1: FUJITSU ATA 3 4GB,master,
/dev/hdc
On the promise 20265 chipset in RAID 0
mode:
ide2:Seagate ATA IV barracuda 80GB, master,
/dev/hde
ide3:Seagate ATA IV barracuda 80GB, master,
/dev/hdg
promise RAID 0 array on
/dev/ataraid/d0
Operating systems and disk setup:
On the FUJITSU, ide1, /dev/hdc I have a redhat 7.2
installed with the 2.4.16 kernel. I have ext2/swap partitions: /(primary), swap
and /home(extented)
On the RAID 0 array, configured with the promise
lite bios 1.31(build 24), I have 4 primary partitions with the GRUB 0.90 as boot
loader.
ataraid/d0p1: windows XP (have to install BEFORE
linux because it complaints that there is no free partition)
araraid/d0p2: reiserfs, / for my
"Own-Linux-I'm-building"
ataraid/d0p3: swap for my
"Own-Linux-I'm-building"
ataraid/d0p4: reiserfs, /home for my
"Own-Linux-I'm-building"
When I use Redhat installed on the fujitsu drive,
it takes forever as soon as there is a disk operation like mount, or
fchk.
In dmesg I have the following (repeated plenty of
times and taking a LOT of time). Some fields varies. I wrote the different
values between ():
hde: read_intr: status=0x59 { DriveReady
SeekComplete DataRequest Error }
hde: read_intr: error=0x10 { SectorIdNotFound },
LBAsect=(184379024, 232590080, 241040279), sector (0,2,4,6)
ide2: reset: success
VFS: Disk change detected on device
ide0(3,0)
For the second line( error=0x10 ), the LBA sector
matches the sector field 0. It's alwas the same LBA sectors. "ide2: reset:
success" seems to appear randomly.
To reproduce the error easily, I use "mount -l"
command line. I tried versions from 2.11g(default in Rehat 7.2) to
2.11n.
Moreover I noticed that if the DMA is enabled for
hde, it will be disabled after such errors(it's printed in dmesg, it is
disabling the DMA on purpose).
The system freezes at the same pulsing rate than
the hard drives are accessed.
Help :(
Sylvain
|
