This is the same issue Steve Cameron just now came and visited me about. If you remember, this was found on a few customer systems running Debian using a 3.x kernel. Steve and I found the cause and the fix was implemented in InfoManager 6.0.14.0 which ultimately got included in the 9.0.13.0 apps release. However, 9.x apps builds have yet to be released publicly until SNAP 1 is wrapped up.
Looks like the version 8.75-x is too early to realize the fix but rest assured that the problem has been addressed.
-Steve D
From: Kilcoyne, William
Sent: Thursday, October 06, 2011 10:13 AM
To: Miller, Mike (OS Dev); Ben Hutchings; Takimoto, Javier
Cc: scameron@beardog.cce.hp.com; 644362@bugs.debian.org; Jonathan Nieder; ISS StorageDev; leo weppelman; Davis, Steven (ISS-Hou); Haselden, Michael
Subject: RE: Bug#644362: linux-image-2.6.39-bpo.2-686-pae: hpacucli hangs when creating disks on cciss
Mike/Steve C/all,
Our code does have logic regarding kernel versions as Steve C mentioned. I have my devs looking into if they have already addressed these concerns.
From: Miller, Mike (OS Dev)
Sent: Thursday, October 06, 2011 9:47 AM
To: Ben Hutchings
Cc: scameron@beardog.cce.hp.com; 644362@bugs.debian.org; Jonathan Nieder; ISS StorageDev; leo weppelman; Kilcoyne, William
Subject: Re: Bug#644362: linux-image-2.6.39-bpo.2-686-pae: hpacucli hangs when creating disks on cciss
On 10/06/2011 09:35 AM, Ben Hutchings wrote:
On Thu, 2011-10-06 at 08:03 -0500, scameron@beardog.cce.hp.com wrote:
On Thu, Oct 06, 2011 at 01:55:37AM -0500, Jonathan Nieder wrote:
leo weppelman wrote[1]:
I tried to reproduce the problem with 3.1.0-rc7-686-pae from
experimental as you requested, but I failed. The problem is that
hpacucli does not find any controllers with this kernel :-( I used the
latest available hpacucli version (8.75-12.0). This version definitely
works with 2.6.39. I tried strace-ing hpacucli but that didn't give me
any clues. It was looking for /dev/cciss/c0d0/cciss, which is weird
and might indicate that the 3.1 environment makes the hpacucli device
search fail....
Weird. Stephen et al: any ideas?
Yeah I suspect I know what's going on. There was some code
in hpaaculi which was asking the question "Is this a 2.4 kernel?
Or is this a 2.6 kernel?" It didn't consider the possibility of
a 3.x kernel, and in the face of a 3.x kernel, it just kind of
gave up.
Pretty sure this bug is fixed in a released version of hpacucli
because rhel6.1 uses a 3.x kernel, and I know we support that.
Bill Kilcoyne might know the specifics.
RHEL 6 will be sticking with what is nominally version 2.6.32, although
RH continues to backport large changes from later versions. The first
stable distribution with a 3.x kernel will presumably be Ubuntu 11.10
'oneiric'.
Stable update 3.0.4 added a process personality flag that causes the
kernel to report version numbers beginning '2.6.', and the 'setarch'
command in util-linux has been updated to allow setting this flag.
However. neither Debian nor Ubuntu has this new version of util-linux
yet.
Ben.
It's sles11 sp2 that uses a 3.0.x kernel.
-- mikem