Hi,
Rafi Sadowsky wrote:
> 2) IIRC correctly there was a decision not to pressure sites to migrate
> from lcg-CE soon
>
> Does that mean that many sites will have to choose between running
> critical nodes with either an unsupported OS(SLC3) or unsupported
> middleware in November ??
>
> Please tell me I'm missing something ....
Yes :-) Putting on my TCG site representative hat now, I can tell you
that at CERN there has already been reasonable success at running the
lcg-CE on a RHEL-4 based OS (in this case SLC4). The TCG has also taken
the strategic decision to capitalize on this effort and make a RHEL-4
lcg-CE an official priority. For exactly the reason you underline above.
It was originally hoped that a replacement CE would be production ready
in time for the SL3 end-of-life but this is not the case.
There is no definitive agreement on how long the RHEL4 lcg-CE will be
supported, but at least until the new CE is production ready; the most
optimistic estimate I have heard is six months.
Hope this helps.
JT
|