ATLAS SCCS Planning 18Jul2007
-----------------------------
9am, SCCS Conf Rm A, to call in +1 510 665 5437, press 1, 3935#
Present: Wei, Richard, Booker, Stephen, Randy, Len
Agenda:
1. DQ2 Status/Web Proxy
Not much information. No too much production. Kaushik is in India
just now so getting jobs is a little slow.
2. Tier-2 Hardware
Depending on where we put it it will depend if we want to optimise
power per rack unit or per watt. Water cooled racks it would be per
rack unit but for Black Box it would be power. Stephen is back at
CERN tomorrow so should hopefully get some numbers for next week.
We know the next purchase will be around $400k with 2/3 storage and
1/3 CPU. About one rack of CPUs. Not sure if this will fit in the
remaining space of the water cooled racks or that space will already
be filled. If the latter happens it would probably go into a second
Black Box.
3. AOB
- Berkeley SRM
Promised that they can provide an SRM gateway next week. Andy
talked to them a month or two ago. They know what we have and what
we'd like. They still need to do some more testing but sounds like
it is nearly ready. Could replace the current gsiftp/posix
interface we currently use.
- End of SLC3
CERN announced this is being end of line in September.
RHEL3 has two or three years left. Issues to do with 32 or 64
bit. ATLAS seems a speed up of 20-30% with 64 bit compiled
applications but twice amount of memory is used.
After nomas and toris are retired all machines will be 64 bit
capable. Should perhaps concentrate on 64-bit way of doing things
as that is the future.
Action Items:
-------------
070711 Stephen Find out about benchmarks for CPUs for next purchase
070718 Not done yet.
