Perhaps because of the attachment it got returned to me too. So here is the content:
rpm -Uvh http://download.fedora.redhat.com/pub/epel/5/x86_64/epel-release-5-4.noarch.rpm
wget http://repository.egi.eu/sw/production/cas/1/current/repo-files/egi-trustanchors.repo
wget http://grid-deployment.web.cern.ch/grid-deployment/glite/repos/3.2/glite-TORQUE_client.repo
wget http://grid-deployment.web.cern.ch/grid-deployment/glite/repos/3.2/glite-GLEXEC_wn.repo
wget http://grid-deployment.web.cern.ch/grid-deployment/glite/repos/3.2/glite-WN.repo
wget http://grid-deployment.web.cern.ch/grid-deployment/download/HEP/repo/HEP_OSlibs.repo
Jeremy
On 3 Aug 2011, at 12:09, wrote:
> Forwarding Raul's response which came to me not the list:
>
>
> one last thing...
>
> dc2-grid-66 has 310 cores (double of that with hyper-threading)
> dgc-grid-43 has 64 cores
>
> Both use torque.
>
> I have noticed in both of them the used memory going up to the point of freezing all gLite services. Either you restart gLite or clean caches periodically.
> I don't know what would happen with EMI Cream plus Torque in a site like RAL or QMUL: thounsands of cores per CE. I believe that Daniela has noticed some memory leaks with her Cream CEs at IC.
>
> Chris: the repos that I used are attached.
>
> Cheers, raul
>
>
> <repos.sh>
>
>
>
> On 3 Aug 2011, at 11:32, <[log in to unmask]>
> wrote:
>
>> Hi Raul,
>>
>> What repos are you using for the EMI CreamCE? We're just replacing one of our remaining lcg-CEs with another CreamCE so I thought we may as well go for the EMI/UMD one.
>>
>> So I installed the UMD repos on the box and tried a yum install emi-cream-ce but that appears to be missing loads of dependencies including things like libglobus* and fetch-crl.
>>
>> Thanks,
>> Chris.
>>
>>> -----Original Message-----
>>> From: Testbed Support for GridPP member institutes [mailto:TB-
>>> [log in to unmask]] On Behalf Of RAUL H C LOPES
>>> Sent: 03 August 2011 11:24
>>> To: [log in to unmask]
>>> Subject: Re: EMI-1 ARGUS testing
>>>
>>> On 02/08/11 21:46, Jeremy Coles wrote:
>>>> Hi Santanu
>>>>
>>>> Raul mentioned to me today that Brunel now have a working (and so far
>>> stable) system with the EMI release (ARGUS and CE). Perhaps you can
>>> check what versions he recommends based on his experience (or maybe
>>> Raul will see this thread and respond before you write and ask!).
>>>>
>>>> Thanks,
>>>> Jeremy
>>>>
>>>>
>>>>
>>>>
>>>> On 2 Aug 2011, at 20:56, Santanu Das wrote:
>>>>
>>>>> On 02/08/2011 20:13, Peter Gronbech wrote:
>>>>>> You should check to see whether it has been approved by egi yet.
>>>>> In that case I can go for the previous version or should I wait?
>>>>>
>>>>> -Santanu
>>>>>> Pete
>>>>>>
>>>>>> ----- Original Message -----
>>>>>> From: Santanu Das [mailto:[log in to unmask]]
>>>>>> Sent: Tuesday, August 02, 2011 06:09 PM
>>>>>> To: [log in to unmask]<[log in to unmask]>
>>>>>> Subject: Re: EMI-1 ARGUS testing
>>>>>>
>>>>>> Hi All,
>>>>>>
>>>>>> I'm planning on installing ARGUS server at Cambridge, so in that
>>> case I
>>>>>> can try EMI-1 ARGUS . Can anyone point me to some documentation
>>> please?
>>>>>>
>>>>>> Cheers,
>>>>>> Santanu
>>>>>>
>>>>>>
>>>>>> On 29/07/2011 11:02, Jeremy Coles wrote:
>>>>>>> Dear All
>>>>>>>
>>>>>>> As mentioned on Tuesday there is a proposal from EMI to "end
>>> standard support for glite 3.2 Argus immediately" in order to focus on
>>> the EMI-1 release. In the UK we have very little experience with the
>>> EMI version and Kashif who (takes part in the staged-rollout tests and
>>> has most knowledge in this area) is away. Brunel does have the EMI
>>> release deployed but also the EMI-CREAM which is the cause of most
>>> current problems they are seeing. Would any other site (using gLite 3.2
>>> CREAM) be prepared to try the EMI-1 ARGUS please? (Or if you happen to
>>> be running it already but did not respond in the meeting on Tuesday
>>> please let us know your experiences).
>>>>>>>
>>>>>>> Many thanks,
>>>>>>> Jeremy
>>> Hi Santanu,
>>>
>>> I have EMI CreamCE at Brunel. It is the Cream CE version 1.13.3
>>> released
>>> last week. I also have an Argus server emi-argus-1.3.0.6.
>>>
>>> Setting up the CreamCE and worker nodes was actually easy. I had
>>> problems with:
>>> - bugs in grid proxy delegation that were only solved last week
>>> - Argus policy: the documentation was not clear in some (at least
>>> from
>>> understanding) and the parser kept silently accepting policies that
>>> were
>>> incorrect.
>>>
>>> Actually I'm still learning about them. For example, in the last few
>>> days the EMI CreamCE started failing the glexec test. The test was
>>> showing as "Warning"
>>> in
>>> https://gridppnagios.physics.ox.ac.uk/nagios/cgi-
>>> bin/status.cgi?host=all. Actually
>>> it was failing with system error 202. Kashif sent me this link
>>>
>>> https://tomtools.cern.ch/confluence/display/SAM/WN#WN-
>>> org.sam.glexec.WNgLExec
>>>
>>> It didn't help much. In the end, I had two problems:
>>> - GLEXEC_LOCATION was not set by yaim, but it is read by the nagios
>>> pilot test
>>> - I had setup debug on in the Argus server one month ago and it
>>> consumed the 65G area that I assigned to logs.
>>>
>>> I do like the Argus authentication, even if I'm a bit scared of the
>>> glexec thing.
>>> I believe things are stable now. Overall I'm very happy with the EMI
>>> CreamCE (dgc-grid-43) and the glite CreamCE (dc2-grid-66) that we
>>> have.
>>>
>>>
>>> Links that I used
>>>
>>>
>>> http://wiki.italiangrid.org/twiki/bin/view/CREAM/SystemAdministratorGui
>>> deForEMI1#1_4_1_Using_the_YAIM_configurati
>>>
>>> https://twiki.cern.ch/twiki/bin/view/EMI/Argus
>>>
>>> https://twiki.cern.ch/twiki/bin/view/LCG/GlexecDeployment#More_informat
>>> ion
>>>
>>> Cheers, raul
>> --
>> Scanned by iCritical.
>
|