[Hallcsw] Sci-Comp meeting notes

[Brad Sawatzky]

Just a couple notes from the last SciComp meeting that may be of
interest:

- There is talk of a another 12 GeV Software Review in September.
  General consensus is that this is a little too close and should be
  pushed back -- it wasn't clear exactly how the September date got set.
  Graham is looking into how much say we have.  Either way, it's time to
  take a look at the promises from last year's review and see how we're
  progressing.

- SciComp will have some significant manpower opening up in the near
  future that Chip would like to apply to "workflow tools" optimized to
  benefit physic's divisions use of the Farm.  Exactly what tools might
  actually be useful will be up to _us_.  He's asked us to think about
  how we go from 'raw data' to 'final result' and identify the slow,
  inconvenient, confusing parts that should be automated, but are not.

  There will be a brainstorming session announced before too long, but
  he asked us to think about what improvements and/or process tools
  might be good for us to have.
  - Options may include:
    - improved Farm job submission process (what will "improved" mean?
      is a web submission interface useful? who knows?)
    - improved end-to-end job monitoring tools?
      - better web front ends?
      - new/improved/simplified failed job and resubmission process?
    - improved file (ie. disk space) management tools?

- We will 'soon' have a formal JLab github account.  This will give us
  full access to all of the github features:
  - bug tracking, wiki, team collaboration tools
  - unlimited public repos
  - 'many' private repos (starting with a medium number, but it will
    grow as we need them)

- The farm migration from RHEL5 to RHEL6/CentOS6 is continuing.  I think
  Hall C is in good shape here, but any groups relying on RHEL5 should
  really be thinking about migrating their code.
  - Right now, there is one Hall B group that is saturating the
    remaining RHEL5 farm nodes.  There is much more RHEL6 horsepower
    available.

- Physics Div is going to be purchasing another disk server (40 TB,
  formatted).  I believe the group is leaning toward adding this to
  /volatile (that would be my recommendation).

-- Brad

-- 
Brad Sawatzky, PhD <brads at jlab.org>  -<>-  Jefferson Lab / Hall C / C111
Ph: 757-269-5947  -<>-  Fax: 757-269-5235  -<>- Pager: brads-page at jlab.org
The most exciting phrase to hear in science, the one that heralds new
  discoveries, is not "Eureka!" but "That's funny..."   -- Isaac Asimov