Tim,
This is exactly what we did at LAL. 2/3 of our workers are in this
configuration and some UIs. I had no time to test other machine types
(except BDII but it is much simpler than other machine types).
I'll update my recipee asap, for gLite3 and send it to this list (previous
post last spring was for LCG 2.7, this is basically the same but a few
details are different). And Markus and Ian suggested to put this on the
wiki.
Michel
--On mardi 10 octobre 2006 22:04 +0200 Tim Bell <[log in to unmask]> wrote:
> During discussions, there has been thoughts on the 32 or 64 bit SLC4
> discussion.
>
> As far as I can see the compatibility options are
>
> 1. SL3 build code on SL4 32 bit
> 2. SL3 build code with appropriate compatibility libraries installed on
> SL4 64 bit
>
> From a pure infrastructure perspective, the 2nd option is very
> interesting since it permits
>
> - infrastructure managers to roll out a single SL4 version (i.e. 64 bits)
> - application code to take advantage of 64 bit if worth while
>
> Will there be a test done on an SL4 64 bit platform with 32 bit
> compatibility libraries ?
>
> Tim
>
> Markus Schulz wrote:
>> Dear Michel,
>> I remember your LCG-2.7 posting.
>> I am currently collecting information on how to run the SL3 build
>> code on SL(C)4 and would be very grateful if you could post an updated
>> version of your HowTo.
>>
>> markus
>>
>> On Oct 10, 2006, at 1:56 PM, Michel Jouvin wrote:
>>
>>> Ilja,
>>>
>>> Anyway, if you want to run current gLite version on SL(C)4, it is
>>> possible and it works very well, including with experiments SW. I
>>> posted a few months ago what you had to do on LCG-ROLLOUT, basically
>>> installing in addition to SL4 (in my case it was 64-bit) a few RPMs
>>> coming from SL3. Probably my post on the list was for LCG 2.7.0 but I
>>> can give you an update for gLite.
>>>
>>> If you are using Quattor (or plan to do it !!!), it is even simpler
>>> as gLite3 on SL4 comes out of the shelf...
>>>
>>> Michel
>>>
>>> --On mardi 10 octobre 2006 14:04 +0300 Ilja Livenson <[log in to unmask]>
>>> wrote:
>>>
>>>> Ok, thanks to all of you, who answered!
>>>> Will be waiting then :)
>>>>
>>>> cheers,
>>>> Ilja
>>>>
>>>> Jeff Templon wrote:
>>>>> Hi Ilja,
>>>>>
>>>>> first a request, please refer to ports to RHEL4 not SLC4. SLC4 is the
>>>>> CERN specific version of Scientific Linux which is a derivative of
>>>>> RedHat Enterprise. If the "porting" efforts -- and even here I would
>>>>> prefer not porting, but "making portable to", I hope that the "ported"
>>>>> version also continues to build and run on RHEL 3.x -- are aimed at
>>>>> RHEL then it is likely that the software runs effortlessly not only on
>>>>> the derivative SLC4, but also other derivatives in use in EGEE/LCG.
>>>>>
>>>>> whew. OK. On to your question: this gets discussed every month at
>>>>> the Grid Deployment Board and other places. You can see e.g. Kors
>>>>> Bos' introductory slides at
>>>>>
>>>>> http://agenda.cern.ch/fullAgenda.php?ida=a057710
>>>>>
>>>>> see slide 8. About the October 31st release, what I recall is that
>>>>> this will be a compatibility mode release, capitalizing on some work
>>>>> done by a couple of Tier-2 sites, enabling the RHEL-3-built software
>>>>> to run in compatibility mode on RHEL-4-based boxes. Someone from SA3
>>>>> will probably pipe up and provide more details at some point.
>>>>>
>>>>> For true RHEL-4-built gLite, my understanding is that this is going to
>>>>> take a while, not so much because of gLite, but because of the
>>>>> underlying VDT distribution, which has never been built on RHEL-4.
>>>>> There are issues with the C++ code, the C++ compiler in RHEL4 is much
>>>>> stricter than that in RHEL3.
>>>>>
>>>>> J "you pick, smiley or sad face on that last one" T
>>>>>
>>>>> Ilja Livenson wrote:
>>>>>> Hello,
>>>>>>
>>>>>> I was wondering - when is the approximate deadline for porting glite
>>>>>> to slc4? So that it would be possible to install middleware without
>>>>>> too much dependency hacks?
>>>>>>
>>>>>> cheers,
>>>>>> Ilja
>>>
>>>
>>>
>>> *************************************************************
>>> * Michel Jouvin Email : [log in to unmask] *
>>> * LAL / CNRS Tel : +33 1 64468932 *
>>> * B.P. 34 Fax : +33 1 69079404 *
>>> * 91898 Orsay Cedex *
>>> * France *
>>> *************************************************************
*************************************************************
* Michel Jouvin Email : [log in to unmask] *
* LAL / CNRS Tel : +33 1 64468932 *
* B.P. 34 Fax : +33 1 69079404 *
* 91898 Orsay Cedex *
* France *
*************************************************************
|