Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
[gentoo-user] Raid system fails to boot after moving from 2.6 kernel to 3.5
19 views
Skip to first unread message
Pau Peris
Oct 26, 2012, 4:40:02 AM10/26/12
to
Hi,
i'm running GNU/Gentoo Linux with a custom compiled kernel and i've just migrated from a 2.6 kernel to a 3.5.
As my HD's are on raid 0 mode i use a custom initrd file in order to be able to boot. While kernel 2.6 is able to boot without problems the new 3.5 compiled kernel fails to boot complaining about "no block devices found". After taking a look at initrd.cpio contained scripts i can see the failure message is given by mdadm tool.
Does anyone has a clue about that? Thansk in advanced. :)
Neil Bothwick
Oct 26, 2012, 5:00:03 AM10/26/12
to
On Fri, 26 Oct 2012 10:36:38 +0200, Pau Peris wrote:
> As my HD's are on raid 0 mode i use a custom initrd file in order to be
> able to boot. While kernel 2.6 is able to boot without problems the new
> 3.5 compiled kernel fails to boot complaining about "no block devices
> found". After taking a look at initrd.cpio contained scripts i can see
> the failure message is given by mdadm tool.
Add "set -x" to the top of your custom init script to see what's going on.
> As my HD's are on raid 0 mode i use a custom initrd file in order to be
> able to boot. While kernel 2.6 is able to boot without problems the new
> 3.5 compiled kernel fails to boot complaining about "no block devices
> found". After taking a look at initrd.cpio contained scripts i can see
> the failure message is given by mdadm tool.
--
Neil Bothwick
Why do programmers get Halloween and Christmas confused?
Because oct 31 is the same as dec 25.
J. Roeleveld
Oct 26, 2012, 5:20:02 AM10/26/12
to
Paul,
I had a similar issue with a new system. There it was caused by mdadm trying to start the raid devices before all the drives were identified.
Try disabling "Asynchronous SCSI scanning" in the kernel config. (CONFIG_SCSI_SCAN_ASYNC)
Or add "scsi_mod.scan=sync" to the kernel commandline to see if it's the same cause.
--
Joost
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.
I had a similar issue with a new system. There it was caused by mdadm trying to start the raid devices before all the drives were identified.
Try disabling "Asynchronous SCSI scanning" in the kernel config. (CONFIG_SCSI_SCAN_ASYNC)
Or add "scsi_mod.scan=sync" to the kernel commandline to see if it's the same cause.
--
Joost
--
Sent from my Android phone with K-9 Mail. Please excuse my brevity.
Pau Peris
Oct 26, 2012, 5:30:02 AM10/26/12
to
Hi,
--
thanks a lot for both answers.
I've just checked my kernel config and CONFIG_SCSI_SCAN_ASYNC is not setted so gonna take a look at it all with "set -x".
Pau Peris Rodriguez
Aquest correu electrònic conté informació de caràcter confidencial dirigida exclusivament al seu/s destinatari/s tant mateix va dirigit exclusivament als seu/s destinatari/s primari/s present/s. Així, queda prohibida la seva divulgació, copia o distribució a tercers sense prèvia autorització escrita per part de Pau Peris Rodriguez. En cas d'haver rebut aquesta informació per error, es demana que es notifiqui immediatament d'aquesta circumstancia mitjançant la direcció electrònica de la persona del emissor.
Paul Hartman
Oct 26, 2012, 11:10:02 AM10/26/12
to
arrays, maybe you using one of the affected versions.
see https://bugs.gentoo.org/show_bug.cgi?id=416081
Pau Peris
Oct 26, 2012, 11:50:02 AM10/26/12
to
Thx a lot Paul,
this morning i noticed there was some kind of issue on my old initrd which works fine for 2.6 kernels, so created a new initrd which works fine and let me to boot into GNU/Gentoo Linux with same 3.5 bzImage.
0 new messages