Re: VIA SATA I/O errors

[Ari Pollak <aripollak@gmail.com>]

Just for the record (in case people see this in the archive and wonder 
what the solution was) - the drive has had a steadily shorter uptime 
over the past few days. When I re-ran WD diagnostics, it finally found a 
problem with the drive, so I'm going to be exchanging it and hopefully a 
new drive will fix the problem.

Ari Pollak wrote:
> Hi.
> I have an Athlon64 machine running kernel 2.6.10-rc3 (but this problem 
> has happened on 2.6.9-ac7 as well) with a VIA VT6420 SATA controller. 
> Every few days (the problem is not chronologically consistent) and/or 
> when there's heavy disk usage,  the main SATA disk (a Western Digital 
> model WDC WD1200JD-00G) will just completely stop responding to any I/O, 
> and a lot of SCSI error messages will be output to the console. After a 
> few instances of this happening (which requires a hard power-off, then 
> power-on.. just hitting the reset button causes the SATA controller not 
> to recognize the drive on boot), I finally managed to capture some of 
> the kernel messages, since somehow I could still read one of my log 
> files (cached in memory, I guess). The same set of errors just keep 
> repeating over and over. I also believe there was an ext3 error that 
> showed up on the console and not in the log, but I assume this is not an 
> ext3 problem anyway. The partial log file and the output of lspci -vvv 
> are attached. I have no idea whether this is a software or hardware 
> problem. Running Western Digital's diagnostics on the drive turned up no 
> errors. If anyone has seen this problem before and it turned out to be 
> hardware-related, I'd like to find out exactly which component is the 
> culprit.
> 
> Thanks in advance,
> Ari