[9fans] Seagate Barracuda 7200.7 write cache?

[9fans] Seagate Barracuda 7200.7 write cache?

[David Eckhardt]

Am I being overly paranoid?

http://www.seagate.com/docs/pdf/datasheet/disc/ds_barracuda7200.7.pdf
is the "data sheet" for the SATA version of the Seagate Barracuda.

Section 4.3.2 ("Set Features command") on logical page 31 (physical
page 41) seems to indicate that the drive defaults to handling sector
writes via a write-back RAM cache, which the host OS can disable.

Is it just me or is this a very bad default except (maybe) for
battery-backed systems?  Is this a standard thing in IDE disks
(either the feature number or the default setting), something
Plan 9 should watch out for, nothing for me to worry my pretty
little head about, ...?

Dave Eckhardt

Re: [9fans] Seagate Barracuda 7200.7 write cache?

[Boris Maryshev]

On Tuesday 13 July 2004 11:25, David Eckhardt wrote:
> Am I being overly paranoid?
>
> http://www.seagate.com/docs/pdf/datasheet/disc/ds_barracuda7200.7.pdf
> is the "data sheet" for the SATA version of the Seagate Barracuda.
>
> Section 4.3.2 ("Set Features command") on logical page 31 (physical
> page 41) seems to indicate that the drive defaults to handling sector
> writes via a write-back RAM cache, which the host OS can disable.
>
> Is it just me or is this a very bad default except (maybe) for
> battery-backed systems?  Is this a standard thing in IDE disks
> (either the feature number or the default setting), something
> Plan 9 should watch out for, nothing for me to worry my pretty
> little head about, ...?
It isn't enabled by default on all IDE HDDs AFAIK. Some recommend to enable
it, because you get increased throughput, but some UNIX vendors do not, as
you can get filesystem corruption in case of power failure. Journalling
doesn't help here either IIRC.
>
> Dave Eckhardt
Boris Maryshev
--
A banker is a fellow who lends you his umbrella when the sun is shining
and wants it back the minute it begins to rain.
		-- Mark Twain

Re: [9fans] Seagate Barracuda 7200.7 write cache?

[David Eckhardt]

In article <200407131148.29808.boris.maroshev@itcollege.ee>,
Boris Maryshev <9fans@cse.psu.edu> wrote:
>It isn't enabled by default on all IDE HDDs AFAIK. Some recommend to enable
>it, because you get increased throughput, but some UNIX vendors do not, as
>you can get filesystem corruption in case of power failure. Journalling
>doesn't help here either IIRC.

Sam Hopkins (sah@softcardsystems.com) pointed out to me in
private mail this charming message:
  http://www.ussg.iu.edu/hypermail/linux/kernel/0304.1/0450.html
which is a post from a Maxtor employee to the effect that IDE
drives are allowed to arbitrarily re-order writes if the "write
cache" is turned on.

Are fossil & venti ok with this?  Should the sd driver try to
turn off the "write cache" feature on IDE drives?

Dave Eckhardt
Followup-To:
Distribution:
Organization: University of Bath Computing Services, UK
Keywords:
Cc:


--
Dennis Davis, BUCS, University of Bath, Bath, BA2 7AY, UK
D.H.Davis@bath.ac.uk

Re: [9fans] Seagate Barracuda 7200.7 write cache?

[ron minnich]

On Fri, 16 Jul 2004, David Eckhardt wrote:

> Sam Hopkins (sah@softcardsystems.com) pointed out to me in
> private mail this charming message:
>   http://www.ussg.iu.edu/hypermail/linux/kernel/0304.1/0450.html
> which is a post from a Maxtor employee to the effect that IDE
> drives are allowed to arbitrarily re-order writes if the "write
> cache" is turned on.


if only the ide drives talked 9p ...

only about 1/4 kidding.

ron

Re: [9fans] Seagate Barracuda 7200.7 write cache?

[Brantley Coile]

[-- Attachment #1: Type: text/plain, Size: 47 bytes --]

Give me time.

  Brantley
  www.coraid.com

[-- Attachment #2: Type: message/rfc822, Size: 3217 bytes --]

From: ron minnich <rminnich@lanl.gov>
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Subject: Re: [9fans] Seagate Barracuda 7200.7 write cache?
Date: Fri, 16 Jul 2004 12:52:18 -0600 (MDT)
Message-ID: <Pine.LNX.4.44.0407161252020.6005-100000@maxroach.lanl.gov>

On Fri, 16 Jul 2004, David Eckhardt wrote:

> Sam Hopkins (sah@softcardsystems.com) pointed out to me in
> private mail this charming message:
>   http://www.ussg.iu.edu/hypermail/linux/kernel/0304.1/0450.html
> which is a post from a Maxtor employee to the effect that IDE
> drives are allowed to arbitrarily re-order writes if the "write
> cache" is turned on.


if only the ide drives talked 9p ...

only about 1/4 kidding.

ron