Hardy config change: Enable CONFIG_NFSD_V4=y in -virtual flavour

[Zack Gilburd]

Tim Gardner wrote:
> Zack Gilburd wrote:
>   
>> Tim Gardner wrote:
>>     
>>> Zack Gilburd wrote:
>>>
>>>  
>>>       
>>>> Forgive my ignorance of the workings of this ML, but I do want to
>>>> point out, regarding the subject of this thread, that this config
>>>> oversight also affects intrepid. I have not checked the jaunty or its
>>>> git, but I would imagine the change would need to be made systemically.
>>>>
>>>> Best regards,
>>>>
>>>> Zack
>>>>
>>>>     
>>>>         
>>> Intrepid doesn't have a -virtual kernel flavour. All Intrepid flavours
>>> already have CONFIG_NFS_V4=y.
>>>
>>>   
>>>       
>> I am sure you are much more in-tune with how exactly these things work
>> than I am.  However, from my observations at $work, the
>> linux-image-virtual in intrepid is based so closely off of the -server
>> flavor that it ships as a separate linux-image, but uses the -server
>> headers.  The linux-image-virtual, as configured, does not have NFSV4
>> client capability.  To test this, you can do a fresh clean JeOS install
>> using an 8.10 CD, dist-upgrade, modprobe nfs, cat /proc/filesystems, or
>> even try to mount NFSV4.  Either way, the problem should become apparent
>> rather quickly.
>>
>> Also worth noting that in intrepid JeOS, /boot/config-`uname -r` appears
>> to be missing.
>>
>> Best regards,
>>
>> Zack
>>
>>     
>
> The virtual kernel in Intrepid _is_ the -server flavour, just slightly
> repackaged. I think the missing /boot/config-`uname -r` is an oversight.
> You can verify the JeOS kernel version by looking at
> /proc/version_signature. I checked and there is an nfsd.ko in
> linux-image-2.6.27-11-virtual_2.6.27-11.23_amd64.deb, so I'm not quite
> sure why you can't make it work. Is this something you've verified at
> runtime?
>
> rtg
>   
Just deployed a fresh image and, as you might have already guessed, 
PEBKAC... My mistake.