[Spacewalk-list] Was: Re: Spacewalk Issue - CentOS kickstarting

John Holland jholland at vizuri.com
Wed Oct 20 20:48:41 UTC 2010


  It would appear I was wrong - I think just creating a new disk file 
got rid of some kind of corruption - I'm getting a good result from a 
virtio disk right now.

On 10/20/2010 02:42 PM, John Holland wrote:
>
>  I just discovered something - maybe everyone on the list already 
> knows - it appears that CentosOS and Fedora guests in KVM booting off 
> of PXE from Spacewalk work a lot better with "IDE" virtual disks than 
> with "Virtio". I'm used to Satellite and RedHat where VirtIO is 
> preferred. Have I reached a correct conclusion, that IDE is better for 
> Spacewalk?
>
> On 10/20/2010 02:17 PM, John Holland wrote:
>>
>>  Thanks,
>>
>> This gets it further along to where it hangs on running pre-install 
>> scripts. (There are no preinstall scripts). Where should I look to 
>> debug this?
>>
>> On 10/20/2010 02:55 AM, Walther, Johannes wrote:
>>> Hi John again,
>>>
>>> I find a solution for me, I meen it is
>>> possible that you do, as test, the same!
>>> CentOS installtion works fine now!
>>>
>>> in other mailings i found:
>>>
>>> ===========================================
>>> Hi all,
>>>
>>>          After spending another day of debugging I think I found the 
>>> problem.
>>>          There seems to be a bug in the kernel (2.6.18-194.el5) 
>>> which was provided for pxe install on the CentOS 5.5 iso images. I 
>>> could trigger it rproducilble via the following steps:
>>>
>>>             * boot it via pxe in rescue mode
>>>             * ensure you have a swap and a date (probably ext3) 
>>> partition
>>>             * try to mount the swap partition ->   this fails
>>>             * try to mount the normal partition ->   system hangs 
>>> ->   kernel panic
>>>
>>>          Can anybody confirm/disprove this?
>>>          @developers: Is there a way to change the kickstart pre 
>>> script which tries to not mount a swap partition?
>>>
>>>          Regards,
>>>          Sascha Bendix
>>>
>>> Sascha,
>>>
>>> Can you try a solution real quick for me?
>>> edit  /var/lib/rhn/kickstart/snippets/keep_system_id
>>>
>>> and change:
>>>      DISKS="$DISKS $(fdisk -l /dev/$disk | awk '/^\/dev/{print $1}')"
>>> to
>>> DISKS="$DISKS $(fdisk -l /dev/$disk | grep -v "swap\|LVM\|Extended" |
>>> awk '/^\/dev/{print $1}')"
>>>
>>> that should ensure that it doesn't try to mount any of those types of
>>> partitions (it seemed that even if it excluded swap but tried to mount
>>> LVM, it would fail).
>>> The above change seemed to work really well for me.
>>>
>>> -Justin Sherrill
>>> ============================================================
>>>
>>>
>>>> -----Ursprüngliche Nachricht-----
>>>> Von: Johannes Walther [mailto:hannes at jw-b.de]
>>>> Gesendet: Mittwoch, 20. Oktober 2010 03:57
>>>> An: John Holland
>>>> Cc: Walther, Johannes
>>>> Betreff: Spacewalk Issue - CentOS kickstarting
>>>>
>>>> Hi John,
>>>>
>>>> Sory, for write you directly, but I cant send messages to the list....
>>>> I dont know why.
>>>>
>>>> I have had the same Problem like you wrote in your Mail to the List.
>>>> -->  [Spacewalk-list] trying to get CentOS 5.5 to kickstart
>>>>
>>>> In Spacewalk I have the channels: CentOS-5.5; Fedora-12;
>>>> Fedora-13; i386
>>>> + x86_64. All with child-channels update; epel; rhn-tools, different
>>>> names.
>>>>
>>>> All fedora  works fine, but boot CentOS, i386 + x86_64, stop
>>>> on the same
>>>> screen like yours. What can i do vor debug this. On the blue-screen
>>>> point the bash (F2) doesn't work.
>>>>
>>>> Cheers and thanks
>>>> Johannes
>>>>
>>>>
>>
>> _______________________________________________
>> Spacewalk-list mailing list
>> Spacewalk-list at redhat.com
>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>
> _______________________________________________
> Spacewalk-list mailing list
> Spacewalk-list at redhat.com
> https://www.redhat.com/mailman/listinfo/spacewalk-list




More information about the Spacewalk-list mailing list