Anyone: Samsung Spinpoint F1 HD753LJ + Si 3512a sata controller problems?

[da_mobius]

Hi,

I've re-installed my Ubuntu box (old Dell poweredge 500sc), and in the process added a SATA pci card with a 3512a chipset on it (cheap brand: Sweex), and two Samsung spinpoint F1 HD753LJ 750GB drives. I configured both drives in a RAID1 configuration. After a succesfull install of the distro without apparent problems, errors started appearing in the log concerning the drives/controller. At first, it only seemed to concern the second drive 'sdb', but last night it also occured with 'sda'.

I also think it's more likely to appear when load on the drives is high, because the errors appear more quickly when I start rebuilding the second drive in order to add it to the RAID array.

Ive tried a number of things, including swapping the drive for a new one, running diagnostic tests on the drive, and 'downgrading' the drives to Sata 1.5 mode, to match the controllers maximum mode.

The errors eventually result in a 'disk failure' error, and subsequently the drive is dropped from the array.

The error typically looks like this (see attachement for a number of these occuring in sequence including the eventual failure, I also included lspci info as an attachement):
[ 155.143914] ata4.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[ 155.143995] ata4.00: BMDMA2 stat 0x650001
[ 155.144132] ata4.00: cmd 35/00:00:bf:02:7a/00:04:00:00:00/e0 tag 0 dma 524288 out
[ 155.144135] res 51/04:19:a6:03:7a/00:03:00:00:00/e0 Emask 0x1 (device error)
[ 155.144314] ata4.00: status: { DRDY ERR }
[ 155.144441] ata4.00: error: { ABRT }
[ 155.178543] ata4.00: configured for UDMA/100
[ 155.178560] ata4: EH complete
[ 185.161846] ata4.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
[ 185.161925] ata4.00: cmd 35/00:00:bf:02:7a/00:04:00:00:00/e0 tag 0 dma 524288 out
[ 185.161928] res 40/00:19:a6:03:7a/00:03:00:00:00/e0 Emask 0x4 (timeout)
[ 185.161985] ata4.00: status: { DRDY }
[ 185.491630] ata4: soft resetting link
[ 190.688791] ata4: port is slow to respond, please be patient (Status 0xd
[ 195.546140] ata4: SRST failed (errno=-16)
[ 195.546198] ata4: hard resetting link
[ 196.055874] ata4: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[ 196.096181] ata4.00: configured for UDMA/100
[ 196.096210] ata4: EH complete
[ 196.106132] sd 3:0:0:0: [sdb] 1465149168 512-byte hardware sectors (750156 MB)
[ 196.114979] sd 3:0:0:0: [sdb] Write Protect is off
[ 196.114985] sd 3:0:0:0: [sdb] Mode Sense: 00 3a 00 00
[ 196.143703] sd 3:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 196.162093] sd 3:0:0:0: [sdb] 1465149168 512-byte hardware sectors (750156 MB)
[ 196.166486] sd 3:0:0:0: [sdb] Write Protect is off
[ 196.166491] sd 3:0:0:0: [sdb] Mode Sense: 00 3a 00 00
[ 196.192033] sd 3:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

Does anyone have an idea where to look for the problem? Is it the unfortunate combination of motherboard/controller/drives that is resulting in this error? Or are the controllers/drives not fully supported?

[mratt]

I have been having this issue for a couple of weeks with my VMWare Server, and by chance tried something which seems to have resolved it for me... well probably now cursed my test server.

I have the following as a test VMWare server
ABIT AB9 Pro which has :
Intel ICH8R with 6 ports
JMicron JMB363 with 2 ports
SIS 3132 with 1int & 1 eSATA
Various drives attached to all and...
Thecus N2050 attached to the SIL 3512 PCI Card that came with it. In it there are 2 WD 250GB drives RAID 1.

Didn't matter what SIS and cable I attached to the Thecus gave the same errors you got.

Something made me :
apt-get install linux-restricted-modules-`uname -r`
In my case : 2.6.24-16-server

Well it hasn't peeped one little bit since the reboot and I have gone much further with tests than in the last 2 weeks. Maybe a module in there that works these chips better. Maybe give it a try if you can.

Regards
MRatt