From mboxrd@z Thu Jan 1 00:00:00 1970 From: Victor Lowther Date: Thu, 16 Jul 2009 07:11:27 +0000 Subject: Re: RFC: mdadm and bringing up raid sets from initrd (dracut) Message-Id: <921e043f0907160011g59e4f75cu788cc7740c91e29e@mail.gmail.com> List-Id: References: <4A5C6501.3080607@redhat.com> <1247581847.1991.16.camel@localhost.localdomain> <20090716001651.GB45537@redhat.com> In-Reply-To: <20090716001651.GB45537-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: Jeremy Katz Cc: Dan Williams , David Zeuthen , Hans de Goede , initramfs , linux-hotplug-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, "Danecki, Jacek" , Harald Hoyer , Doug Ledford , NeilBrown On Wed, Jul 15, 2009 at 7:16 PM, Jeremy Katz wrote: > On Wednesday, July 15 2009, Dan Williams said: >> mdadm-3.0 has facilities to prevent assembly of certain metadata types >> [1] or arrays with certain uuids [2]. =A0I wonder if we also need a >> facility to prevent auto-assembly of arrays *not* listed in >> mdadm.conf? =A0So the mdadm.conf file installed in the initramfs would >> only identify the root array and all other randomly identified md >> devices would be ignored (rather than assembled with a foreign name). >> >> Thoughts? > > There is no mdadm.conf in the initramfs -- in fact, the initramfs may > not even be generated on the system that you're booting and instead be > "generic" for the kernel in question Still, it sounds like a good feature to be added for a --hostonly initramfs. (especially on systems that are attaching to iscsi and/or fibre channel lun= s) > Jeremy > -- > To unsubscribe from this list: send the line "unsubscribe initramfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at =A0http://vger.kernel.org/majordomo-info.html >