> -----Original Message-----
> From: Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Henry Nebrensky
> Sent: 06 October 2005 14:29
> To: [log in to unmask]
> Subject: Re: Reminder of the next UKI meeting - Wednesday 5th
> October 11:00-12:00 via VRVS
>
> We all discussed it at an LT2 mtg way back - the Atlas issue
> prompted it but an explicit consideration was that we
> wouldn't try to back-up user data on SEs, and instead put the
> saved resources (time, money for tape drives, etc.) back
> into, er, resources (CPU and disk).
My point is that this needs to be a wider discussion. If LT2 ends up
offering a service level that no-one wants it is a waste of many things
including the discussion. We need to match experiment requirements with
service provision. One might think this would have been done pre-MoU but
obviously not down to this level.
> In particular, permanent meant that once the data was written
> in it would remain available (until deleted by the VO).
> This has consequences both in terms of VOs not losing data
> completely, but also in classic SEs being required to stay
> running for posterity, as data may need to be accessed via a
> TURL/SFN held in a non-LCG catalog (which can't thus be
> automatically updated when moving to an SRM).
I think you are reading too much into 'permanent'. You can always close
down or rename an SE but you have to do it together with the VO. For the
reasons you mention, they have to move or rename/recatalogue data.
>
>
> If experiments are happy that for random disk-based SEs
> around the Grid (rather than a small, specific set of tape
> silos) "permanent" can in fact mean
> "permanent-except-some-risk-of-hardware-failure" then the
> immediate issue mostly disappears (thus remaining unsolved
> until the next crisis :( ).
Exactly
>
> Of course, as Tier 2 SEs are there to act as local data
> caches rather than long-term storage, making them
> volatile/durable doesn't seem conceptually unreasonable.
No, as long as that is what the experiments want and they have the
middleware to handle it.
|