linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Daniel Brahneborg <basic@wtnord.net>
To: "Måns Rullgård" <mru@kth.se>
Cc: linux-raid@vger.kernel.org, linux-ide@vger.kernel.org
Subject: Re: RAID5 over Serial-ATA success stories?
Date: Thu, 8 Jan 2004 08:57:11 +0100	[thread overview]
Message-ID: <20040108085711.A2500@nettis.grimsta> (raw)
In-Reply-To: <yw1xn08z8ltb.fsf@ford.guide>; from mru@kth.se on Wed, Jan 07, 2004 at 10:56:32PM +0100

Thanks for the feedback, it's very valuable to me.

On Wed, Jan 07, 2004 at 10:56:32PM +0100, Måns Rullgård wrote:
> Daniel Brahneborg <daniel.com@wtnord.net> writes:
> 
> > I'd like to hear some success stories for RAID5 on Serial-ATA disks.
> > Which Serial-ATA card are you using?  Do you get decent performance?
> > Is it stable with DMA enabled?  Do you use the 2.4 or 2.6 kernel?
> >
> > I know this much:
> >
> > It doesn't work with Silicon Image.
> 
> What doesn't work?  There are drivers, at least in 2.6.  Raid should
> care about what sort of disks you use.

When using it for normal disks e2fsck reports bad blocks all
over the disk. When used for RAID, I get corrupted data.  Not
much, maybe every second time for a file of 500MB.
This is with the IDE driver.  With the SCSI driver, my computer
completely freezes when I activate my second network card (as I
reported earlier, unfortunately still without a solution).
RAID might work with that driver, but unless the network card
problem is solved, that doesn't help me.

> > It doesn't work with VIA (yet, anyway).
> > It might work with HighPoint.
> 
> I've run RAID5 on a Highpoint RocketRAID 1540.  I used ATA disks with
> SATA converters, though.  Works with both 2.4 and 2.6.

Sounds good to hear.  It's the second cheapest card for me.

> > It probably works with Promise.
> > I don't know if there's a driver for Adaptec.
> 
> Which Adaptec card?  The 12xx cards are fakeraid, but are supported as
> normal cards.  The 24xx cards are true hardware RAID cards.  Linux
> drivers exist for these, too.

It's the 12xx cards that I'm looking at. I don't want hardware
RAID, since hardware RAID5 costs an infinite amount of money.

> > In case I have to replace my Silicon Image card, what should I replace
> > it with?  I'm currently leaning towards Promise TX4 (or TX2 if the VIA
> > driver is completed).
> 
> I stay as far away as I can from Promise and VIA.  Anything is usually
> better than those two.

Why the warning about Promise?  The reason I want the VIA
driver to work is that I've got two VIA connectors on the
motherboard, so I only need a 2 port SATA card.

/Basic


  reply	other threads:[~2004-01-08  7:57 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-01-07 21:36 RAID5 over Serial-ATA success stories? Daniel Brahneborg
2004-01-07 21:56 ` Måns Rullgård
2004-01-08  7:57   ` Daniel Brahneborg [this message]
2004-01-08  8:06     ` Måns Rullgård
2004-01-08  8:27       ` Daniel Brahneborg
2004-01-08 11:48     ` A.J.Dawson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20040108085711.A2500@nettis.grimsta \
    --to=basic@wtnord.net \
    --cc=linux-ide@vger.kernel.org \
    --cc=linux-raid@vger.kernel.org \
    --cc=mru@kth.se \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).