Question for Neil/Superblocks

Question for Neil/Superblocks

[Justin Piszcz]

Hi Neil,

A couple of things, recently I tried to make a new system (MDADM/latest/SW 
RAID1) and noticed that GRUB & LILO still only support 0.90 superblocks. 
This leads me to my next question.

Will the newer superblocks be supported? I ask this due to my final 
question.

With 0.90 superblocks, I have never seen any array (raid1,raid5,raid6) try 
to correct a read-error, but when I made an array with 1.2 superblocks 
(not the boot volume of course) using VelociRaptors, great for testing 
crap disks that don't work in read, I did see it try to correct read 
errors!  I was wondering if that is only applicable to superblocks >= 
1.x+?

I know you can build yourself an initrd/miniroot so/on/so forth but its 
nice when you can just build a kernel, run lilo and have it auto-detect 
the raid1 volumes and boot accordingly (as you can with 0.90 superblocks).

Justin.

Re: Question for Neil/Superblocks

[Kasper Sandberg]

On Fri, 2009-11-27 at 16:27 -0500, Justin Piszcz wrote:
> Hi Neil,
> 
> A couple of things, recently I tried to make a new system (MDADM/latest/SW 
> RAID1) and noticed that GRUB & LILO still only support 0.90 superblocks. 
> This leads me to my next question.
> 
> Will the newer superblocks be supported? I ask this due to my final 
> question.
> 
> With 0.90 superblocks, I have never seen any array (raid1,raid5,raid6) try 
> to correct a read-error, but when I made an array with 1.2 superblocks 
> (not the boot volume of course) using VelociRaptors, great for testing 
> crap disks that don't work in read, I did see it try to correct read 
> errors!  I was wondering if that is only applicable to superblocks >= 
> 1.x+?
> 
> I know you can build yourself an initrd/miniroot so/on/so forth but its 
> nice when you can just build a kernel, run lilo and have it auto-detect 
> the raid1 volumes and boot accordingly (as you can with 0.90 superblocks).
have a separate boot partition raid1....


> 
> Justin.
> --
> 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

RE: Question for Neil/Superblocks

[Simon Jackson]

> With 0.90 superblocks, I have never seen any array (raid1,raid5,raid6) try 
> to correct a read-error, but when I made an array with 1.2 superblocks 
> (not the boot volume of course) using VelociRaptors, great for testing 
> crap disks that don't work in read, I did see it try to correct read 
> errors!  I was wondering if that is only applicable to superblocks >= 
> 1.x+?

Is this actually true that RAID 1 volumes with 0.90 Superblock will not do read recovery?



-----Original Message-----
From: linux-raid-owner@vger.kernel.org [mailto:linux-raid-owner@vger.kernel.org] On Behalf Of Kasper Sandberg
Sent: 27 November 2009 22:22
To: Justin Piszcz
Cc: linux-raid
Subject: Re: Question for Neil/Superblocks

On Fri, 2009-11-27 at 16:27 -0500, Justin Piszcz wrote:
> Hi Neil,
> 
> A couple of things, recently I tried to make a new system (MDADM/latest/SW 
> RAID1) and noticed that GRUB & LILO still only support 0.90 superblocks. 
> This leads me to my next question.
> 
> Will the newer superblocks be supported? I ask this due to my final 
> question.
> 
> With 0.90 superblocks, I have never seen any array (raid1,raid5,raid6) try 
> to correct a read-error, but when I made an array with 1.2 superblocks 
> (not the boot volume of course) using VelociRaptors, great for testing 
> crap disks that don't work in read, I did see it try to correct read 
> errors!  I was wondering if that is only applicable to superblocks >= 
> 1.x+?
> 
> I know you can build yourself an initrd/miniroot so/on/so forth but its 
> nice when you can just build a kernel, run lilo and have it auto-detect 
> the raid1 volumes and boot accordingly (as you can with 0.90 superblocks).
have a separate boot partition raid1....


> 
> Justin.
> --
> 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

--
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

Re: Question for Neil/Superblocks

[Doug Ledford]

[-- Attachment #1: Type: text/plain, Size: 1466 bytes --]

On 11/27/2009 04:27 PM, Justin Piszcz wrote:
> Hi Neil,
> 
> A couple of things, recently I tried to make a new system
> (MDADM/latest/SW RAID1) and noticed that GRUB & LILO still only support
> 0.90 superblocks. This leads me to my next question.
> 
> Will the newer superblocks be supported? I ask this due to my final
> question.

Not by grub or lilo, which are both dead projects.  However, grub2 does
support newer superblocks and we actually ship it in F12.

> With 0.90 superblocks, I have never seen any array (raid1,raid5,raid6)
> try to correct a read-error, but when I made an array with 1.2
> superblocks (not the boot volume of course) using VelociRaptors, great
> for testing crap disks that don't work in read, I did see it try to
> correct read errors!  I was wondering if that is only applicable to
> superblocks >= 1.x+?

Coincidence.

> I know you can build yourself an initrd/miniroot so/on/so forth but its
> nice when you can just build a kernel, run lilo and have it auto-detect
> the raid1 volumes and boot accordingly (as you can with 0.90 superblocks).

The hack in lilo that allows it to do this is a very ugly hack.  It is
not even close to proper raid1 boot support.  I would look into grub2.

-- 
Doug Ledford <dledford@redhat.com>
              GPG KeyID: CFBFF194
	      http://people.redhat.com/dledford

Infiniband specific RPMs available at
	      http://people.redhat.com/dledford/Infiniband


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]