|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|
Re: [PLUG] Linux (Debian) and raid
|
Hi
On Mon, 2007-05-28 at 08:22 +0200, sean finney wrote:
> On Sunday 27 May 2007 23:58:18 Antony P Joseph wrote:
> > Is it reasonable to put boot image in the raid there by causing
> > unnecessary requirements on the boot loader?
> > 1. A small boot loader has to recognize the raid etc..
> > Boot image is used only once at boot time before the next boot.
There are two ways I think we can incorporate this into grub. One is to
incorporate as stage 1.5 raid loader. Second is to add alternate or
secondary boot in case the primary fails(no raid). (This is just a
guess)
>
> the way lilo works it doesn't actually put the bootloader on the md device at
> all, but instead puts it on each of the physical disks' boot sections. so
> the configuration could be thought of as "and put one here too".
>
> > If you are booting the device again, you have plenty of time to put
> > another backup drive of boot device in place.
>
> this makes a very large assumption that you know your disk has failed. think
> about what happens if the first time it fails is when you turn on your
> computer :)
Probably I did not explain clearly.
There is no assumption here. My point is that boot image is read only
and used only once at boot time. After that a failure of the disk or
partition containing boot image is irrelevant or even noticeable to the
running of the system until next boot. You can detect that failure only
at the next boot time assuming you do not have some fancy disk
monitoring software. Usually in a well run shop, booting of the server
is well planned in advance . You will have sufficient time to replace
the failed disk with new backup disk, CDROM boot or even a USB stick
with the boot image.
Regarding the mirroring of the data.
If I want mirror the data, usually I will do a separation based on the
type of data. For example operating system(except boot image)on mirrored
disks and Mysql in another mirrored disks. If some catastrophe happens
such as double disk failure, controller running amok or user application
running amok, I will be burdened with salvaging the data of the both
application( the OS and MySql ) at the same time. My point here is that
you can not just walk away saying that I created one big partition(IT IS
SO EASY), mirrored the partition in another disk and everything is fault
tolerant now.
With regards
Antony
>
>
> sean
> ___________________________________________________________________________
> Philadelphia Linux Users Group -- http://www.phillylinux.org
> Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
> General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug
___________________________________________________________________________
Philadelphia Linux Users Group -- http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug
|
|