SATA-problems
Heya,
I've been having some strange SATA-problems and I'm not sure what's
'causing them. I noticed first about a week ago when XMMS just froze for
no reason at all. Checking the log (/var/log/messages) revealed a lot of
messages looking like this:
Jul 4 12:32:11 grimreaper kernel: ata2: command 0xc8 timeout, stat 0xd0
host_stat 0x1
Jul 4 12:32:11 grimreaper kernel: ata2: translated ATA stat/err 0xd0/00
to SCSI SK/ASC/ASCQ 0xb/47/00
Jul 4 12:32:11 grimreaper kernel: ata2: status=0xd0 { Busy }
Jul 4 12:32:11 grimreaper kernel: sd 1:0:0:0: SCSI error: return code =
0x8000002
Jul 4 12:32:11 grimreaper kernel: sda: Current: sense key: Aborted Command
Jul 4 12:32:11 grimreaper kernel: Additional sense: Scsi parity error
Jul 4 12:32:11 grimreaper kernel: Info fld=0x25cc897
Jul 4 12:32:11 grimreaper kernel: end_request: I/O error, dev sda,
sector 39635095
Anything depending on files on the drive froze so I issued a reboot, but
that didn't work either. I had to turn the machine off and then start it
up again to get it working. Since then it's happened a few times and
it's all random (at least I've not seen any pattern). I'm using a
PATA-drive for the root-system and the SATA-drive is only used for
/home. The PATA-drive is using ReiserFS and the SATA-drive is using ext3
as filesystems.
My hardware configuration is as follows:
- Fortron/Source PSU ATX 350W
- AMD Athlon 64 3000+ 1.8GHz Socket 939, 512KB, BOXED
- MSI K8N NEO4-FI,nForce4 Ultra,Socket-939 Raid, Firewire, SATAII,
GbLAN, PCI-Ex16
- Corsair Value S. PC3200 DDR-DIMM 2048MB Kit w/two matched Value Select
1024MB
- MSI GeForce 7600GS 256MB DDRII, PCI-Express, NX7600GS-T2D256EH, DVI-I
- Samsung SpinPoint P120S 250GB SATA2 8MB 7200RPM NCQ
- Western Digital Caviar 120GB IDE 7200RPM Special Edition 8MBcache
- NEC ND-2500A DVD-burner
I'm running Debian (sid/unstable) with the 2.6.17.3 kernel. I was
running 2.6.16.18 when the problem first appeared and have also tried
2.6.16.21, 2.6.17 and 2.6.17.1. I ran a full scan with Memtest86+ just
to rule the RAM out and then a full scan with Samsungs diagnostics tool
(HUTIL) which didn't find anything wrong (it ran a full surface scan
amongst other things).
I thought that perhaps it's some bug in the sata_nv driver so I tried
hooking up my old SATA-controller card (Sunsway/ST Lab PCI SATA 2P,
SiL3112), but alas it was not the case (I'm still running on the SiL3112
card).
I've put up some information on http://web.telia.com/~u85920559/sata/
including:
- .config for 2.6.17.3
- dmesg output for 2.6.17.3
- syslog errors messages (04-jul and 07-jul)
- smartctl -H -A output (04-jul and 07-jul)
Any help is greatly appreciated.
Thanks in advance,
Fredrik
Reply to: