> How does one "mark" a new CE (fronting a WN SL6 queue) as SL6 WN? Is
it just assumed? Who needs to be notified?
In general the BDII has a GlueHostOperatingSystem* with Name, Release,
Version
In the specific:
* LHCb uses the BDII to find new SL6 queues
* Atlas you need to notify atlas in UK [log in to unmask]
because they need to validate the releases. If you have a new queue
they'll have to add it to the system. However since it is a pain - I'm
currently doing this for Oxford and Imperial - you are encourage to use
existing queues already in the system. Even in this case the releases
still need to be validated.
* CMS/Alice have usually local contacts who know what to do.
All this is in the SL6Migration
(https://twiki.cern.ch/twiki/bin/view/LCG/SL6Migration#Procedures_and_how_to_contact_ex)
page which everyone is encouraged to read.
cheers
alessandra
On 14/05/2013 16:13, Stephen Jones wrote:
> Hi there;
>
> here's a few questions for TB_SUPPORTers to chew on.
>
> How does one "mark" a new CE (fronting a WN SL6 queue) as SL6 WN? Is
> it just assumed? Who needs to be notified?
>
> Should one use a different tag directory? Or a different software
> directory? Or CVMFS repos etc?
>
> Is it worth supporting WN-SL6 in an EMI2 CREAM/TORQUE/WN stack, or is
> it best to jump straight to EMI3?
>
> What do you (anyone) think? Any tips gratefully received.
>
>
> Cheers,
>
> Steve
>
>
--
Facts aren't facts if they come from the wrong people. (Paul Krugman)
|