From: Duncan <1i5t5.duncan@cox.net>
To: gentoo-amd64@lists.gentoo.org
Subject: [gentoo-amd64] Re: Root on Raid and LVM
Date: Mon, 31 Oct 2005 01:50:33 -0700 [thread overview]
Message-ID: <pan.2005.10.31.08.50.32.496956@cox.net> (raw)
In-Reply-To: 433FDF0A.4090306@gmx.at
Florian D. posted <433FDF0A.4090306@gmx.at>, excerpted below, on Sun, 02
Oct 2005 15:22:18 +0200:
> i cannot understand, how you got things working without an extra /boot
> partition. the system needs initrd to start the LVM, right? but if
> initrd itself is on the LVM...? very interesting ;-)
He didn't say he didn't have an extra boot. He only enumerated the LVs
(logical volumes) on the MD/RAID, which he said was on sda5, sdb5, etc,
all partition 5, on /dev/md5.
The usual way to handle that, if you are booting from the RAID array
as well, is to create a small raid1 array out of small partitions toward
the front of each physical disk, for /boot. Being raid1 (mirrored), the
same data is imaged to all disks in the raid1 identically. Both GRUB and
LILO can read individual disk partitions belonging to the mirror array as
if they were stand-alone partitions, so can boot from any of them, altho
they do boot from only one at a time. The kernel and initrd are on this
boot array (so mirrored to each physical disk in the array), and can then
be loaded. Root and other LVs are then created on the main VG (volume
group (or multiple groups) taking the rest of the disks, to be loaded from
the initrd. Once /they/ are loaded, and the pivot_root to the main-root
on the main VG is accomplished, the raid1 array containing /boot can be
mounted, if desired. (As usual, Gentoo keeps /boot unmounted by default.)
... As some may guess, I've been studying this stuff recently! =8^) I
don't have my own RAID setup yet, but probably will by late this week. (I
plan to go pickup the drives probably Tue or Wed.)
Of course, he didn't actually say /boot is on raid1, either. It may be on
a conventional disk, perhaps manually backed up. There is some clue that
there are extra partitions, however, and probably extra RAID volumes as
well, since both the partitions and the md's mentioned were #5. Now, #5
is the first secondary partition in the extended partition, so there
wouldn't /have/ to be a 1-4. However, there's likely at least an sdX1 on
all the disks, which would form the raid1 for /boot.
--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html
--
gentoo-amd64@gentoo.org mailing list
next prev parent reply other threads:[~2005-10-31 8:55 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-10-02 1:00 [gentoo-amd64] Root on Raid and LVM Chris S
2005-10-02 2:47 ` Chris Smart
2005-10-02 3:17 ` [gentoo-amd64] Root on Raid and LVM - Solved Chris Smart
2005-10-02 13:22 ` [gentoo-amd64] Root on Raid and LVM Florian D.
2005-10-31 8:50 ` Duncan [this message]
2005-11-02 12:33 ` [gentoo-amd64] " Florian D.
2005-11-02 12:44 ` Sebastian Redl
2005-11-02 22:12 ` [gentoo-amd64] " Duncan
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=pan.2005.10.31.08.50.32.496956@cox.net \
--to=1i5t5.duncan@cox.net \
--cc=gentoo-amd64@lists.gentoo.org \
/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