Re: ahci errors on 8-stable

[Harald Schmalzbauer]

Nenhum_de_Nos schrieb am 16.03.2010 02:01 (localtime):

On Mon, March 15, 2010 14:21, Harald Schmalzbauer wrote:

Nenhum_de_Nos schrieb am 09.03.2010 00:44 (localtime):

On Mon, 8 Mar 2010 14:26:53 -0800
Jeremy Chadwick <free...@jdc.parodius.com> wrote:

On Mon, Mar 08, 2010 at 06:38:02PM -0300, Nenhum_de_Nos wrote:

I've seen these errors in a production machine in deep disk load (scp
and
bsdtar in heavy use):

Please provide the output from the following commands:

As I had huge disk activity when those messages appeared, I did reboot
after and now no more are there. I think the vmstat command should be
issued when the problem was happening right ? (if so I can run the
backup tar's and see what happens).

What disks do you use?
I have similar timeouts and mav has the hd firmware in mind to be the
culprit
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/stable/2010-02/msg00737.html

In my case it's the samsung EcoGreen SpinPouint F2 1.5TB, Firmware
1AG01113 and 1AG01118. The disk on ahcich2 (where the timeouts appear)
has the newer firmware.

2 Seagate 1TB disks:

Mar  8 13:49:45 optimus kernel: ada1 at ahcich1 bus 0 scbus1 target 0 lun 0
Mar  8 13:49:45 optimus kernel: ada1: <ST31000528AS CC38> ATA-8 SATA 2.x
device
Mar  8 13:49:45 optimus kernel: ada1: 300.000MB/s transfers (SATA 2.x,
UDMA6, PIO size 8192bytes)
Mar  8 13:49:45 optimus kernel: ada1: Command Queueing enabled
Mar  8 13:49:45 optimus kernel: ada1: 953869MB (1953525168 512 byte
sectors: 16H 63S/T 16383C)
Mar  8 13:49:45 optimus kernel: ada2 at ahcich3 bus 0 scbus3 target 0 lun 0
Mar  8 13:49:45 optimus kernel: ada2: <ST31000528AS CC38> ATA-8 SATA 2.x
device
Mar  8 13:49:45 optimus kernel: ada2: 300.000MB/s transfers (SATA 2.x,
UDMA6, PIO size 8192bytes)
Mar  8 13:49:45 optimus kernel: ada2: Command Queueing enabled
Mar  8 13:49:45 optimus kernel: ada2: 953869MB (1953525168 512 byte
sectors: 16H 63S/T 16383C)

those are known to be bad ?

Thanks,

-Harry

In my experience, these are reliable drives. And completely different to mine. So I think it's not liklely to be a firmware bug. I hope the problem can be pointed out. If there's anything I can help, please let me know.

signature.asc
Description: OpenPGP digital signature