Yep ... if we are going to take something away, we at least need to give 
them some warning.  Putting it back in was rather trivial, it was all 
the rebuilding and retesting everything that is/was the bigger pain.
On the last test right now ...

Troy

Urs Beyerle wrote:
> 
> Troy Dawson wrote:
>> OK, this will just drive me nuts if I don't fix it.
>> Fixing
>> I am putting xfs-filesystem along with the kernel-modules into i386. 
>> I am also changing the comps file so that if someone picks xfs, it
>> will automatically pull in xfs-filesystem, just like on SL 5.3.
> 
> 
> I think this is a good decision! to avoid some frustrated users.
> 
> Cheers,
> 
>     Urs
> 
> 
> 
> 
>> Troy
>>
>> Troy J Dawson wrote:
>>> Hi Urs,
>>> You are correct, and this is something I didn't notice.  My desktop
>>> is a 64 bit machine, I saw XFS as a kernel module, and I assumed that
>>> it was a kernel module in 32 bit as well.
>>>
>>> We have always been telling people to only put XFS on 64 bit
>>> machines, but it looks like RedHat is doing more than just telling
>>> people.
>>>
>>> So the question is, what do we do?
>>> For security updates, I guess we have to provide i386 kernel module
>>> rpm's still.
>>>
>>> For the SL 5.4 release ... well ... you can't actually do anything
>>> during the install with xfs.  So, is it that bad that the kernel
>>> module isn't in there?
>>>
>>> But part of me wonders if it isn't better to just put the needed
>>> stuff in, rebuild everything, retest, and get it out.  It isn't that
>>> dramatic of a change, and if this was monday, I wouldn't even think
>>> about it.
>>> But part of me knows the dangers of XFS on 32 bit to begin with ...
>>> so I'm a bit torn.
>>>
>>> Troy
>>>
>>> Urs Beyerle wrote:
>>>> Hi Troy and Connie,
>>>>
>>>> I'm a little bit confused about xfs support.
>>>> (Sorry, but I have no time to check it myself.)
>>>>
>>>> Did we have kernel-module-xfs for i368 in SL5.3.? I think yes.
>>>> Now in SL5.4 xfs module is in the x86_64 kernel, but not in the i368
>>>> kernel, right?
>>>> If this is true, it will make xfs formated partition unusable after an
>>>> update to SL5.4 (for i386)...?
>>>>
>>>> Cheers,
>>>>
>>>>     Urs
>>>>
>>>>
>>>>
>>>> Troy Dawson wrote:
>>>>> Release Candidate 2.5 for Scientific Linux 5.4 has been released
>>>>> for both i386 and x86_64.
>>>>>
>>>>> This is essentially the release.  All the release notes say it is the
>>>>> release, everything is in the correct place, and we don't plan on
>>>>> changing anything unless there is a real show stopper.
>>>>>
>>>>> So this is your final chance to test.  If there isn't a real show
>>>>> stopper, then tomorrow we will change the 5x link to point to this,
>>>>> and we will officially announce the release of Scientific Linux 5.4
>>>>>
>>>>> -----------------------------------------------------------------------------
>>>>>
>>>>>
>>>>> DOWNLOAD INFO
>>>>> -----------------------------------------------------------------------------
>>>>>
>>>>>
>>>>>
>>>>>       http://ftp.scientificlinux.org/linux/scientific/5rolling/i386
>>>>>       http://ftp.scientificlinux.org/linux/scientific/5rolling/x86_64
>>>>>       http://ftp.scientificlinux.org/linux/scientific/5rolling/iso
>>>>> or
>>>>>       http://ftp.scientificlinux.org/linux/scientific/54/i386
>>>>>       http://ftp.scientificlinux.org/linux/scientific/54/x86_64
>>>>>       http://ftp.scientificlinux.org/linux/scientific/54/iso
>>>>>
>>>>> -----------------------------------------------------------------------------
>>>>>
>>>>>
>>>>> CHANGES by Scientific Linux
>>>>> -----------------------------------------------------------------------------
>>>>>
>>>>>
>>>>>
>>>>> ** Added all the latest errata and fastbugs
>>>>>
>>>>> ** Verified signatures of rpm's
>>>>> **  (Note: jdk x86_64 is a known exception)
>>>>>
>>>>> ** Verified rpm versions, releases, and names, fixing those that were
>>>>> not correct
>>>>>
>>>>> ** CD and DVD install boot screen now says what version and arch you
>>>>> are installing
>>>>>
>>>>> ** General Cleanup of release notes, readme files and other minor
>>>>> stuff that doesn't really affect the release, but makes things look
>>>>> better.
>>>
>>


-- 
__________________________________________________
Troy Dawson  [log in to unmask]  (630)840-6468
Fermilab  ComputingDivision/LSCS/CSI/USS Group
__________________________________________________