Hi,
The so called "central" lfc is prod-lfc-atlas-central.cern.ch. It
is not used by the ATLAS DDM but unfortunately has a legacy role
for the users.
I get confused with acronyms but the indico meeting title I
referred to is WLCG-OSG-EGEE Operations meeting
We thought the hostname change was an internal site issue. We
were aware of several hard-wired configurations around the world,
e.g. FTS and LFC, and if the change exposed more then that`s a
good thing.
In an ideal world there would be no hard-wired places to change,
and thanks to the infosys, it almost worked. The 2 problems I
know of FTS config - the design was to use the infosys but the
BDII was not reliable enough so the services.xml file was born.
LFC SFN - the schema has the hostname buried in the SFN, for
reasons I still can`t fathom. If the hostname was a foreign key
to a table of SE:hostname mappings then the change would be
trivial. At 300,000 entries, it`s still fairly trivial with a
'replace'.
As I said, in the end the change became urgent due the hardware
problems on the old headnode. Apologies if it went too fast for
anyone, but better now than during data-taking.
Cheers,
Rod.
On Mon, 5 Feb 2007, Jamie Shiers wrote:
>
> I do not believe that it has been done on the central LFC whatever
> that is.
>
> But the real point is I am not convinced that we ever agreed
that this
> was the correct procedure.
>
> There is, BTW, no WLCG operations meeting. We have a joint
operations
> meeting, plus an LCG Experiment Coordination Meeting and a Service
> Coordination meeting. It has not been brought up in either the
latter
> two of these.
>
> I will put it on the agenda for today's LCG ECM.
>
> Cheers,
>
>
--
Tel. +1 604 222 7667