opensuse
[Arriba] [Todas las Listas]

Re: [opensuse] Re: REDADA de software vs REDADA de BIOS

To: opensuse@xxxxxxxxxxxx
Subject: Re: [opensuse] Re: REDADA de software vs REDADA de BIOS
From: "Brian K. White" <brian@xxxxxxxxx>
Date: Mon, 19 Sep 2011 01:35:11 -0400
Delivered-to: opensuse@xxxxxxxxxxxxxxxxxxx
Delivery-date: Mon, 19 Sep 2011 01:35:35 -0400
Envelope-to: traductor@xxxxxxxxxxx
In-reply-to: <4E75E16C.1030905@rogers.com>
List-archive: <http://lists.opensuse.org/opensuse/>
List-help: <mailto:opensuse+help@opensuse.org>
List-owner: <mailto:opensuse+owner@opensuse.org>
List-post: <mailto:opensuse@opensuse.org>
List-subscribe: <mailto:opensuse+subscribe@opensuse.org>
List-unsubscribe: <mailto:opensuse+unsubscribe@opensuse.org>
Mailing-list: contact opensuse+help@xxxxxxxxxxxx; run by mlmmj
Organization: Aljex Software Inc.
References: <4E6FFFBE.102@gmail.com> <4E70011E.5010101@gmail.com> <j4phqn$qpp$2@saturn.local.net> <4E71CE83.8030702__39131.0333543972$1316081301$gmane$org@rogers.com> <j50pha$qt1$1@dough.gmane.org> <4E75E16C.1030905@rogers.com>
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:6.0.2) Gecko/20110902 Thunderbird/6.0.2
En 9/18/2011 8:17 AM, James *Knott escribió:
*Joachim *Schrod escribió:
Aquí *LVM es el problema, no REDADA-1.

Con REDADA-1 *via *md y ningún *LVM, /bota en un *mirrored *ext3/*ext4
*filesystem definitivamente trabajos con *grub. Lo utilizamos desde varios
años, y tuvo bastantes errores de disco para probarlo laborable. Uno necesita cuidar que *grub *bootloader es instalado en ambos discos, aun así,
*openSUSE  no  esto él durante instalación. La Robustez
extra viene con *stanzas que nombra ambos discos, en caso *hd0 no
es visible al BIOS más y *hd1 de repente consigue *hd0.


Quizás alguien tendría que hacer un *HOWTO en esto. Yo recientemente montado un *server en
trabajo con REDADA 4 puesto que la mayoría de el sistema, pero /la bota no es en REDADA. Si
había sabido sobre esto lo habría hecho.


El problema es hay varios pasos que son esencialmente *unsupportable por *suse hasta que construyen el apoyo a *yast y otros guiones de sistema y herramientas.

Lo adivino sería fresco *anyways pero sería un *longish documento porque tendrías que explicar todo que el sistema no automáticamente.

Entonces otra vez tal *howtos ha existido para otro *distros para años
*http://www.texsoft.it/índice.*php?*c=Hardware&*m=*hw.*storage.*grubraid1&*l=él

De hecho es sido *doable para tan mucho tiempo que muchos del existiendo *howto referir a *lilo en vez de *grub o referir a construir vuestro *kernel para conseguir el apoyo. No necesitamos ninguno de aquello.

Necesitaríamos documentar:
* Cómo a a mano reunir el principal no-*bootable variedad si el *initrd fallado para hacerlo, el cual pasa a veces. * Cómo para asegurar *initrd consigue regenerado con los módulos de redada requeridos incluyeron, porque por *default lo poder o poder no. (Edita /*etc/*sysconfig/*kernel y entonces corrido *mkinitrd, y quizás hacer todo esto después de a mano chutando al instalar *media, a mano reuniendo la variedad, montando el *filesystem, liga montar /*dev/ /*proc /*sys al montado *fs, y finalmente *chrooting al instalado pero no todavía *bootable sistema...)

--
*bkw
--
A *unsubscribe, *e-correo: *opensuse+unsubscribe@xxxxxxxxxxxx
Puesto que órdenes adicionales, *e-correo: *opensuse+help@xxxxxxxxxxxx


On 9/18/2011 8:17 AM, James Knott wrote:
Joachim Schrod wrote:
Here LVM is the problem, not RAID-1.

With RAID-1 via md and no LVM, /boot on a mirrored ext3/ext4
filesystem definitively works with grub. We use it since several
years, and had enough disk errors to prove it working. One needs to
take care that grub bootloader is installed on both disks, though,
openSUSE doesn't do this itself during installation. Extra
robustness comes with stanzas that name both disks, in case hd0 is
not visible to the BIOS any more and hd1 suddenly gets hd0.


Perhaps someone should do a HOWTO on this. I recently set up a server at
work with RAID 4 for most of the system, but /boot is not on RAID. If
I'd known about this I would have done it.


The problem is there are several steps that are essentially unsupportable by suse until they build the support into yast and other system scripts and tools.

I guess it would be cool anyways but it would be a longish document because you'd have to explain everything that the system doesn't do automatically.

Then again such howtos have existed for other distros for years
http://www.texsoft.it/index.php?c=hardware&m=hw.storage.grubraid1&l=it

In fact it's been doable for so long that many of the existing howto's refer to lilo instead of grub or refer to building your kernel to get the support. We don't need any of that.

We WOULD need to document:
* How to manually assemble the main non-bootable array if the initrd failed to do it, which does happen sometimes. * How to ensure initrd gets regenerated with the required raid modules included, because by default it may or may not. (edit /etc/sysconfig/kernel and then run mkinitrd, and maybe do all this after manually booting into the install media, manually assembling the array, mounting the filesystem, bind mounting /dev/ /proc /sys into the mounted fs, and finally chrooting into the installed but not yet bootable system...)

--
bkw
--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse+help@xxxxxxxxxxxx


<Anterior por Tema] Tema Actual [Siguiente por Tema>