Print

Print


ATLAS SCCS Planning 12Dec2007
-----------------------------

  9am, SCCS Conf Rm A, to call in +1 510 665 5437, press 1, 3935#

Present: Wei, Stephen, Richard, Randy, Chuck, JohnB, JohnW, Len

Agenda:

1. DQ2 Status/Web Proxy

    Production is running well. Using 300 CPUs most of the time. Have
    enough data to run jobs. Have FUSE running on a few xrootd
    machines. Also have a Berkeley version of SRM (BeStMan), transferred
    data from SLAC to UTA using FTS. This is called BeStManXroot but it
    doesn't use any features of xrootd so it can be without it.

    For the dq2_put problem, it might be better to use noric09 which
    has XrootdFS running.

    Do we know if the fairshare works correctly for ATLAS? We believe
    so. There are users using some of it so it doesn't all go to
    production.

2. Tier-2 Hardware

    Asked to replicate 12/13 release AOD. Would need about 40TB, don't
    have enough space for this. Spending ATLAS money on disk would get
    us about 200TB. We pushed ahead with the CPU spending as it was
    desired. Can now get the 1TB disks so that will make disk pricing
    more attractive. Currently in negotiation about storage
    price. BaBar can probably wait till April for the next big
    purchase. Alexei suggested getting AOD since 1st September. Not
    sure how useful a definition that is but can meanwhile do that and
    determine if there are requests for other data. Currently importing
    1 or 2TB of RDO data for trigger work.

    Order for CPUs went out about a week ago. Expect delivery before
    Christmas but installation after the break. Will probably move some
    machines from water cooled racks to Sun Black Box. The full set of
    machines should be available early February. Existing ATLAS
    machines will move to the second Black Box (it'll be half BaBar,
    quarter ATLAS - all current installation - and quarter GLAST).



3. tcsh with larger environment space

    Taylor needs to copy it across to the machines. It'll go to
    /usr/etc/tcsh and put that /etc/shells. Then users should ask for
    their shell to be changed. Will need to be changed on Solaris too
    to ensure you can have the same shell everywhere.

4. Tier2/3 Workshop

    Not much to say, it appears on the ATLAS e-News today.

5. AOB

    - Space for atl-prod02 mirror

    Reply from Wei's ticket saying there wasn't a 72GB one. Will have
    another, perhaps larger one.

Action Items:
-------------

071114	Stephen Attempt to use tcsh 6.14 to see if that solves env problem
 	071212 Done. Will deploy.