Re: [PROBLEM] reproduceable storage errors on high IO load

["Lars Täuber" <taeuber@bbaw.de>]

[-- Attachment #1: Type: text/plain, Size: 757 bytes --]

Hi!

The memory test with memtest86+ v4.10 has passed 8 times without
finding any error. (64 hours for 4GB RAM with Chipkill ECC)

I made some more tests running vanilla 2.6.39.2.
The box is mainly idle only doing the io to the disks.

1.
* Ultrastar 7K3000 attached to onboard AMD_AHCI
* dd if=/dev/zero of=/dev/sde
* dd told write speed of: 15.7MB/s
* no error after half an hour

2.
* Ultrastar 7K3000 attached to PCIe to sata_mv controller
* dd if=/dev/zero of=/dev/sda
* dd told write speed of: 6.5MB/s
* no error but strange log: dmesg-PCIe-sata_mv.log.gz

3.
* Ultrastar 7K3000 attached to PCIe areca SAS HBA (mvsas)
* dd if=/dev/zero of=/dev/sdg
* dd told write speed of: 1.3MB/s
* errors: dmesg-PCIe-mvsas.log.gz

Has someone an idea?

Thanks
Lars

[-- Attachment #2: dmesg-PCIe-sata_mv.log.gz --]
[-- Type: application/octet-stream, Size: 16581 bytes --]

[-- Attachment #3: dmesg-PCIe-mvsas.log.gz --]
[-- Type: application/octet-stream, Size: 23353 bytes --]