From: Ken Gunderson <kgunders@teamcool.net>
To: NeilBrown <neilb@suse.de>
Cc: linux-raid@vger.kernel.org
Subject: Re: non fs-data and gpt partitioned md
Date: Wed, 18 Apr 2012 19:02:21 -0600 [thread overview]
Message-ID: <20120418190221.561b0df6.kgunders@teamcool.net> (raw)
In-Reply-To: <20120419103823.30ca2834@notabene.brown>
On Thu, 19 Apr 2012 10:38:23 +1000
NeilBrown <neilb@suse.de> wrote:
> On Wed, 18 Apr 2012 18:18:33 -0600 Ken Gunderson
> <kgunders@teamcool.net> wrote:
>
> > On Tue, 17 Apr 2012 14:40:10 -0600
> > Ken Gunderson <kgunders@teamcool.net> wrote:
> >
> > Apologies for following my own post but I guess should elaborate...
> >
> > > Hello:
> > >
> > > I'm wanting to set up a new md root lvm based configuration and
> > > after reading various docs am confused about how I should be
> > > going about this. My intent is to have mirrored /boot and swap
> > > partitions and a raid10 / partition with LVM.
> > >
> > > Issue #1: gpt is recommended over mbr based partitioning for new
> > > installs, even on BIOS based systems (presuming these BIOS will
> > > boot gpt, wh/mine do). Auto-assemble is not recommended for new
> > > installations (my understanding is that it is not necessary with
> > > initramfs) so I presume I should be specifying partition type DA.
> > > However, while gdisk et.al. allow for selecting type FD, they do
> > > not offer DA as an option.
> >
> > Per <http://www.rodsbooks.com/linux-fs-code.html>, Linux ignores
> > partition type codes. Yet per man mdadm:
> >
> > "When creating a partition based array, using mdadm with
> > version-1.x metadata, the partition type should be set to 0xDA (non
> > fs-data). This type selection allows for greater precision since
> > using any other [RAID auto-detect (0xFD) or a GNU/Linux partition
> > (0x83)], might create prob‐ lems in the event of array recovery
> > through a live cdrom."
> >
> > So which is it? Does partition type code matter to md/mdadm or not?
>
> Both. Neither.
>
> md does handle 0xFD partitions a bit differently, but I recommend not
> using that feature.
> Other than that md igores them. mdadm ingores them completely.
>
> But other tools - typically installers - might pay some attention to
> them. Using 0xDA discourages such tools from mishandling them.
Thanks for the clarification.
> > > Issue #2: Is there any reason to prefer 1.0 vs. 1.2 metadata? I
> > > can use either grub2 or Syslinux boot loaders. My understanding
> > > is that Syslinux supports the former while Grub2 supports 1.2.
> > > All other things being equal, I'd prefer to use Syslinux. Unless
> > > there is some technical reason to favor 1.2 metadata and/or Grub2.
> > >
> > > So what would be best practices recommended way to proceed here?
> >
> > The reason I ask is that I bring such a configuration online on
> > Archlinux by following these instructions:
> >
> > <https://wiki.archlinux.org/index.php/Software_RAID_and_LVM>
> >
> > But I'll be damned if I can recover from failed drive simulations -
> > at least reliably, as sometimes it works while others not - so I'm
> > just trying to rule out potential variables here.
> >
> > Thanks-- Ken
> >
>
> Any reason for preferring one of 1.0 and 1.2 is out side of md.
>
> Maybe you want to be able to mount one half of a RAID1
> independently. You need 1.0 for that. Maybe you want to ensure that
> never happens. Then 1.2 is better. Maybe your boot loaded only works
> with one. Then the choice is clear.
Well, this only pertains to /boot partition, so there's nothing to stop
one from mixing and matching, e.g. Syslinux using 1.0 metadata for
RAID1 /boot and 1.2 metadata for RAID10 swap and root devices, no? Or
is this not advisable?
Thanks-- Ken
--
Ken Gunderson <kgunders@teamcool.net>
--
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:[~2012-04-19 1:02 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-04-17 20:40 non fs-data and gpt partitioned md Ken Gunderson
2012-04-19 0:18 ` Ken Gunderson
2012-04-19 0:38 ` NeilBrown
2012-04-19 1:02 ` Ken Gunderson [this message]
2012-04-19 1:36 ` NeilBrown
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=20120418190221.561b0df6.kgunders@teamcool.net \
--to=kgunders@teamcool.net \
--cc=linux-raid@vger.kernel.org \
--cc=neilb@suse.de \
/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).