lkml.org 
[lkml]   [2002]   [Jul]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: IBM Desktar disk problem?
On Fri, 05 Jul 2002, Thunder from the hill wrote:

> ...and tell all the people who got a DTLA (because it's not as expensive
> as others in some countries, mind France, USA, Germany) to drop their
> disks if they want to use Linux, because we're too lazy to find a
> solution. That might be cool to you, but we want HARDWARE SUPPORT for
> Linux! That's why we're here.
>
> There _is_ a solution, we just have to find it.

Might just be that "coincidence" happens more often as "incident"
happens more often. Means: the more drives of that type fails, the more
likely it is someone shows up with a report that the failure is related
to this-and-that. Until someone can reproduce this, assume the TCQ is
Not Guilty in respect to your DTLA failure. Assume it has died a natural
death.

And the "incident often" seems to hold for 75GXP series drives (IBM
DTLA-3070xx). Out of seven DTLA-307045 (all bought approx. Feb. 2001,
in two different stores) I look after I am getting a MTBF(*) of around
28,000 hours -- which is way lower than what IBM claims. Two died after
9 months, another one after 16 months. Other IBM drives, DJNA, DPTA, are
way more reliable -- but these are no longer sold.

(Søren Schmidt, FreeBSD ata(4) driver maintainer, says that IBM
acknowledged that DJNA have broken TCQ, sometimes forget to flush a
block -- this will then probably also affect the WDC-420400 and 418....
20G 5400/min and 18G 7200/min drives).

(*) Using Fujitsu's formula:

total hours of service of all drives
MTBF = --------------------------------------
number of failed drives to date[unhandled content-type:application/pgp-signature]
\
 
 \ /
  Last update: 2005-03-22 13:27    [from the cache]
©2003-2011 Jasper Spaans