kernel panic on a 2950

[Kaj Niemi]

Hi,


The multiple scsi_hostadapter0 entries are redundant, one is enough.

Assuming the failed machine and the other are identical in hardware  
you should be able to copy over modprobe.conf and then regenerate the  
initrd image as per below. If you're running from the rescue disk you  
obviously need to change the paths (I'm kind of assuming you are).

1. copy the current /boot/initrd-kernelversion.ELsmp.img to something  
else just in case

2. /sbin/new-kernel-pkg -v --package kernel --mkinitrd --depmod -- 
install 2.6.9-67.0.15.0.1.EL (do rpm -q --scripts kernel-yourversion  
or kernel-smp-yourversion to see the correct syntax, -v is just for  
verbosity)

Should you want to look into the current initrd image, it's typically  
a gzipped cpio archive (verify with "file") so gunzipping it first and  
then "cpio -ivd":ing into a temporary directory will help you out  
there. It could also be interesting to compare the contents of the  
init file from the original initrd and the one you generated after  
running new-kernel-pkg.



tx

Kaj

On May 14, 2008, at 18:29, Chamberland, Robert wrote:

>
>
>> -----Original Message-----
>> From: David_Kewley at Dell.com [mailto:David_Kewley at Dell.com]
>> Sent: Tuesday, May 13, 2008 5:52 PM
>> To: Chamberland, Robert
>> Cc: linux-poweredge at lists.us.dell.com
>> Subject: RE: kernel panic on a 2950
>>
>>> -----Original Message-----
>>> From: linux-poweredge-bounces at dell.com [mailto:linux-poweredge-
>>> bounces at dell.com] On Behalf Of Kaj Niemi
>>> Sent: Tuesday, May 13, 2008 11:21 AM
>>>
>>> On May 13, 2008, at 20:31, Chamberland, Robert wrote:
>>>
>>>> Just before the RHEL rescue disk kicks in the controller correctly
>>>> reports seeing two logical drives.
>>>> The rescue disk reports not seeing the drives and asks if I want
> to
>>>> install drivers for them (which I skip).
>>>
>>> Did you, by any chance, happen to boot with another kernel than the
>>> one you were running before rebooting? If the VDs aren't seen it
> could
>>> be that the initrd image does not contain the necessary drivers to
>>> talk with the card; this usually means mkinitrd got hosed when
>>> installing the new kernel.
>>
>> Following up on Kaj's idea:
>>
>> If you have an original 2950, the built-in RAID controller is PERC
> 5/i,
>> and its driver is megaraid_sas.  You need a line like this in
>> /etc/modprobe.conf:
>>
>> alias scsi_hostadapter megaraid_sas
>>
>> (If you have other scsi_hostadapter* lines, it might need to be
>> scsi_hostadapterN, where N is 1, 2, 3, 4,...)
>>
>> Is it possible this line got removed when you were messing with NIC
>> configs?  If so, and if you manually or automatically updated your
>> initrd (automatic update could happen via a kernel or kernel module
> rpm
>> install/update), then the new initrd would not have the driver.  When
> a
>> boot is attempted using this initrd, it would not be able to locate
> your
>> hardware RAID, which means it could not access /.  This would match
> your
>> original kernel panic symptoms.
>>
>> Do you have another kernel in GRUB that you can boot into?  Maybe its
>> initrd has the necessary drivers.
>>
>> David
>>
>>
>> David Kewley
>
>
> You and Kaj are onto something.  No, I didn't load a new kernel and
> don't have options for another kernel but /etc/modprobe.conf isn't
> loading megaraid_sas.
>
> This is the modprobe.conf from my good machine:
>
> alias scsi_hostadapter megaraid_sas
> alias usb-controller ehci-hcd
> alias usb-controller1 uhci-hcd
> alias scsi_hostadapter0 mptsas
> alias scsi_hostadapter0 mptsas
> alias scsi_hostadapter0 mptsas
> alias scsi_hostadapter0 mptsas
> alias eth2 e1000
> alias eth3 e1000
> alias eth0 bnx2
> alias eth1 bnx2
>
> This is from my bad machine:
>
> alias eth2 e1000
> alias eth3 e1000
> alias usb-controller ehci-hcd
> alias usb-controller1 uhci-hcd
> alias eth1 bnx2
> alias eth0 bnx2
>
> And when I was messing with ifconfig my changes weren't taking,  
> which is
> probably interesting in this context...hrm.
>
> So would the solution be to simply use the modprobe.conf from the good
> machine?
>
> Rob C
>
> _______________________________________________
> Linux-PowerEdge mailing list
> Linux-PowerEdge at dell.com
> http://lists.us.dell.com/mailman/listinfo/linux-poweredge
> Please read the FAQ at http://lists.us.dell.com/faq




HTH

Kaj
-- 
Kaj J. Niemi
<kajtzu at basen.net>
+358 45 63 12000



-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3811 bytes
Desc: not available
Url : http://lists.us.dell.com/pipermail/linux-poweredge/attachments/20080514/be2f769e/attachment.p7s