[et-mgmt-tools] Re: strange virt-install/koan issue

[Michael DeHaan]

Brenton Leanhardt wrote:
> +++ Michael DeHaan [16/07/08 13:45 -0400]:
>>
>>> I suspect that if we can get koan to not set that
>>> attribute it will solve my problem.
>>>>
>>>> _______________________________________________
>>>> et-mgmt-tools mailing list
>>>> et-mgmt-tools at redhat.com
>>>> https://www.redhat.com/mailman/listinfo/et-mgmt-tools
>>> ------------------------------------------------------------------------ 
>>>
>>>
>>> _______________________________________________
>>> et-mgmt-tools mailing list
>>> et-mgmt-tools at redhat.com
>>> https://www.redhat.com/mailman/listinfo/et-mgmt-tools
>>
>> Here is what koan does in qcreate.py:
>>
>> if arch is not None and arch.lower() == "x86":
>> arch = "i686"
>>
>> guest = 
>> virtinst.FullVirtGuest(hypervisorURI="qemu:///system",type=vtype, 
>> arch=arch)
>>
>> I am not sure of the ramifications of leaving off the arch or why 
>> that might be breaking things. Perhaps remove that arch bit in koan 
>> and try it?
>>
>> --Michael
>
> Ha, how simple. That does indeed work for me. If that's an
> acceptible solution I'd love to see it make it in into a release.
>
> I'm still puzzled as to why a guest arch of 'i386' doesn't work on my
> hardware while 'i686' does.
>
> --Brenton
>>
>>
>>
>> _______________________________________________
>> et-mgmt-tools mailing list
>> et-mgmt-tools at redhat.com
>> https://www.redhat.com/mailman/listinfo/et-mgmt-tools

This is there because cobbler uses i386 as an internal representation to 
say "generically intel but not x86_64", though the code is set to
send what libvirt wants to libvirt. IIRC, yum wants to see i386, so it's 
just glue there.

--Michael