Re: device timeouts on hard disks no matter what..

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Mon, Dec 07, 2009 at 08:58:56PM +0200, Teemu Rinta-aho wrote:
> Hi all,
> 
> I started to got these device timeouts while writing or reading
> hard disks. I thought it was my cheap Mini-ITX motherboard that
> was too slow or something, but now that I have changed everything
> else than NetBSD, and I just got this, basically without any
> load on the server:
> 
> wd1a: device timeout reading fsbn 301867968 of 301867968-301868095 (wd1 bn 
> 301868031; cn 299472 tn 4 sn 3), retrying
> ahcisata1 port 1: device present, speed: 3.0Gb/s
> wd1: soft error (corrected)
> wd1a: device timeout reading fsbn 521875008 of 521875008-521875135 (wd1 bn 
> 521875071; cn 517733 tn 3 sn 18), retrying
> ahcisata1 port 1: device present, speed: 3.0Gb/s
> wd1: soft error (corrected)
> 
> I emphasize that I have first changed disks, then motherboard, cables,
> chassis, and the rest, except the operating system.
> 
> Why are these errors there? Do you get them? Is it certain that it is
> a hardware problem or would there be something to be done in the NetBSD
> code? Do all operating systems get those but they simply don't
> print anything into their logs??

Linux at last also print them (but with a different format of course).
I'm not seeing this on production machines. I've seen this on occasion,
and it turned out to be hardware-related (the disk is dying and takes too
long retrying to get the data, or the power supply is too short).
If you already tried new disks, I would try a stronger power supply next
(this fixed the issue for me on several occasion).

What model of controller is it ? I see it's called ahcisata1
so I guess it's a add-on ?

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--