From: "Joël Bourquard" <numlock@freesurf.ch>
To: Mark Hahn <hahn@physics.mcmaster.ca>
Cc: mingz@ele.uri.edu, admin@ecommerce.com, linux-raid@vger.kernel.org
Subject: Re: What SATA controller for software raid and/or lvm?
Date: Mon, 19 Jan 2004 20:50:07 +0100 [thread overview]
Message-ID: <1074541807.11253.33.camel@localhost.localdomain> (raw)
In-Reply-To: <Pine.LNX.4.44.0401191411050.12676-100000@coffee.psychology.mcmaster.ca>
Mark,
Thanks a lot for your answer...
On Mon, 2004-01-19 at 20:19, Mark Hahn wrote:
> > However, before I rush and order these two Highpoint 1820's ($300), it
>
> afaikt, HPT's driver is one that plugs into SCSI. this by itself doesn't
> mean it's bad (since libata and 3ware also do so, and are good). the problem
> is that it's not clear how well HPT will maintain the driver, updating it
> when the SCSI layers above it change (or the kernel in general).
>
I understand your concerns about the future !
Fortunately (at least for me) it won't matter much, because I don't need
top-notch software in the fileserver.
Since we talk SCSI, is it still possible to monitor the S.M.A.R.T health
attributes of the SATA disks when plugged into one of these controllers?
I'd be lost with 16 disks and no smartctl..
By the way, there's something I didn't grasp yet.. is libata a library
which is needed for SATA drivers to run ? Or does libata rather mean
"open-source SATA-to-SCSI driver" ?
In the second case, with a HPT controller there would be no need for
libata, correct ?
> I'm going with promise 4-port cards: most server boards have two PCIX
> buses, it gets painful to put more than 8 disks in a chassis anyway.
> these days, a disk will do 70 MB/s, so expecting 560 MB/s out of a single
> card is probably over-optimistic anyway.
>
Good point. With 8 disks, I'd have taken two 4-port controllers too.
> > it might help a lot of people on Linux-RAID... I mean, a 64-bit 133MHz
> > SATA controller with 8 ports, for about $150 is really nice :-)
>
> I don't mind two cheaper cards for the same thing. it's easy to argue
> that you shouldn't have the controller as a single point of failure, too.
>
Indeed. Also if there are two arrays (I think it's needed because of the
2TB limit), one could split the work on the 2 controllers... and on the
two PCI-X bridges :-)
> regards, mark hahn.
Regards
Joël
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2004-01-19 19:50 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-01-18 20:54 What SATA controller for software raid and/or lvm? Ming Zhang
2004-01-18 21:29 ` Hendrik Visage
2004-01-18 21:51 ` Joël Bourquard
2004-01-18 22:29 ` Ming Zhang
2004-01-18 23:50 ` Joël Bourquard
2004-01-19 19:19 ` Mark Hahn
2004-01-19 19:50 ` Joël Bourquard [this message]
2004-01-19 23:16 ` Mark Hahn
2004-01-18 22:26 ` Ming Zhang
2004-01-18 23:02 ` Hendrik Visage
2004-01-19 0:22 ` Ming Zhang
2004-01-19 19:10 ` Mark Hahn
2004-01-19 19:27 ` Hendrik Visage
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=1074541807.11253.33.camel@localhost.localdomain \
--to=numlock@freesurf.ch \
--cc=admin@ecommerce.com \
--cc=hahn@physics.mcmaster.ca \
--cc=linux-raid@vger.kernel.org \
--cc=mingz@ele.uri.edu \
/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