PS if only slc5 and slc6 are foreseen SL(-C) sites might have a similar problem.

cheers
alessandra

 On 07/05/2013 10:48, Alessandra Forti wrote:

Well Atlas is aware of the issue I don't know when the fix will be available.

From the Rod entry it seems cmt reads the architecture from some OS file and the CentOS architecture is not foreseen. Either the CeontOS sites will have to hack it and replace the architecture or atlas has to add a CentOS arch. The sw librarian is CC'd in the ticket and now also here.

cheers
alessandra


On 07/05/2013 10:36, Daniela Bauer wrote:

Hi Alessandra,

I was more hoping for a 'Atlas is aware of the issue, a fix is forthcoming' as it seems to suggest an Atlas specific problem.

The ticket is almost incomprehensible (can someone fix those html junk characters ?)

Cheers,

Daniela

On 7 May 2013 10:29, Alessandra Forti <[log in to unmask]> wrote:

Hi Daniela,

Rod updated the ticket with a question yesterday. You might want to subscribe the ticket.

cheers
alessandra

On 07/05/2013 10:25, Daniela Bauer wrote:

Hi,

Is there any input from Atlas on this ?

Cheers,

Daniela

On 6 May 2013 09:54, Rodney Walker <[log in to unmask]> wrote:

Hi,

I know the LRZ fix, but it is not nice. I will add something useful to the ticket.

Cheers,

Rod.

On 6 May 2013 10:23, Paco Bernabe <[log in to unmask]> wrote:

Hi everybody,

A GGUS ticket (https://ggus.eu/ws/ticket_info.php?ticket=93242) to our site (SARA-MATRIX) regarding a known issue about atlas analysis jobs landing on SL6 WNs described in the wiki (https://twiki.cern.ch/twiki/bin/view/Atlas/RPMCompatSLC6#compiling_packages_against_SLC5). When upgrading our WNs to CentOS6, we read the whole wiki and performed the necessary actions to avoid issues, but this code we cannot change (Meaning that it is not a site responsibility):

	# for building slc5 binaries on slc6 host
	macro CppSpecificFlags "" \
	    ATLAS&host-slc6&target-slc5 " -D__USE_XOPEN2K8 "

So it looks that the macro doesn't get defined for CentOS6 sites, but only slc6 sites (At least at our site doesn't). Now, the user that opened the ticket sent us a link to another GGUS ticket (https://ggus.eu/ws/ticket_info.php?ticket=93271) of a site (LRZ-LMU) that had the same problem and a workaround was found; this site has installed SLES11 SP2 which is compatible to SL6. If anybody of LRZ-LMU gets this email, could you let us know about that workaround? Is there any other site that support ATLAS and with SL6 (Or compatible) that has also this issue? Please, let us know.

--

Met vriendelijke groeten / Best regards,

Paco Bernabe

| Systemsprogrammer | SURFsara | Science Park 140 | 1098XG Amsterdam | T +31 610 961 785 | [log in to unmask] | www.surfsara.nl |

--
Tel. +49 89 289 14152

--

Sent from the pit of despair

-----------------------------------------------------------
[log in to unmask]
HEP Group/Physics Dep
Imperial College
London, SW7 2BW
Tel: +44-(0)20-75947810
http://www.hep.ph.ic.ac.uk/~dbauer/

-- 
Facts aren't facts if they come from the wrong people. (Paul Krugman)

--

Sent from the pit of despair

-----------------------------------------------------------
[log in to unmask]
HEP Group/Physics Dep
Imperial College
London, SW7 2BW
Tel: +44-(0)20-75947810
http://www.hep.ph.ic.ac.uk/~dbauer/

-- 
Facts aren't facts if they come from the wrong people. (Paul Krugman)

-- 
Facts aren't facts if they come from the wrong people. (Paul Krugman)