From: "NeilBrown" <neilb@suse.de>
To: "mark delfman" <markdelfman@googlemail.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: RAID60 - MD / LVM
Date: Thu, 23 Jul 2009 09:06:07 +1000 (EST) [thread overview]
Message-ID: <bde080c5e5b50b0cb99492dd561190bd.squirrel@neil.brown.name> (raw)
In-Reply-To: <66781b10907220204o26964a70j90348ec78a685b52@mail.gmail.com>
On Wed, July 22, 2009 7:04 pm, mark delfman wrote:
> Hi All,
>
> Firstly let me apologise for what maybe a simple question, I have
> tried to google and search this out.
>
> The question is simply the best way to create a RAID60 or any nested
> RAID. Looking at the examples it seems the approach is to create 2 x
> RAID6s and then simply RAID0 on top. However, is there any advantage
> in using MD to create the RAID0 or would you see the same overall
> result by using LVM to stripe (R0) the 2 x RAID6s?
Most of the difference between using MD and LVM to do the striping would
be simply in the fact that different tools are used, so the best advice
would probably be to use whatever you are more familiar with.
In 2.6.31, md/raid0 will support a chunk size that is not a power of
2. I don't think LVM supports that (yet).
That was added specifically for this sort of set up, so that the
RAID0 chuck can align with the RAID6 stripe. Exactly how important
that alignment is probably depends on your workload. I haven't really
thought it though enough to be able to give any guidance.
>
> Given that I see a lot of posts RE: RAID10, then I suspect that the
> MD RAID0 code has some understanding of or tuning for nested RAIDs,
> which I guess LVM may not have.
MD has a separate RAID10 module which stripes multiple copies of the
data across multiple drive. It is simpler and more flexible than
RAID0 over RAID1. There is no similar RAID60 module as there is a lot
less room for gains in flexibility.
>
> It would seem the answer is obvious as it makes sense to simply use
> MD, but we have an internal system based on 100s of old scripts and
> the RAID section only allows us to select drives (not raids), it
> does however use LVM so it would be easier to change the script which
> created the VG as opposed to change the many scripts related to RAID
> creation...
I guess that is your answer then - use LVM over MD.
>
> If there is a real disadvantage in using LVM, then we can manually
> create the RAIDs I guess....
I haven't used LVM in this configuration so I cannot report first
hand experience, but I would be very surprised if LVM did not meet
your needs.
NeilBrown
>
>
>
> I would appreciate any feedback
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
next prev parent reply other threads:[~2009-07-22 23:06 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-07-22 9:04 mark delfman
2009-07-22 23:06 ` NeilBrown [this message]
2009-07-26 16:04 ` mark delfman
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=bde080c5e5b50b0cb99492dd561190bd.squirrel@neil.brown.name \
--to=neilb@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=markdelfman@googlemail.com \
/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
all inboxes | Powered by JetHome®