r2 - 20 Feb 2010 - 04:28:41 - RichardMountYou are here: TWiki >  AtlasSoftware Web > Minutes12Feb2010

Minutes12Feb2010 - RAC Minutes, February 12, 2010


Members (*=present)

*Richard Mount, *Kevin Black (Physics Forum Chair), *Jim Cochran (Analysis Support Manager), Alexei Klimentov (ATLAS ADC) [not available at this meeting time] Ian Hinchliffe (Physics Advisor), *Rik Yoshida (Tier3 Coordinator) *Michael Ernst (U.S. Facilities Manager) *Rob Gardner (Integration Coordinator) *Kaushik De (U.S. Operations Manager) Armen Vartapetian (U.S. Operations Deputy)

Ex-Officio: * Torre Wenaus * Stephane Willocq [first 15 minutes] * Mike Tuts * Howard Gordon

Brief Introductions from Each Member

Many members roles were obvious.

Kevin/Jim/Ian have roles in delivering physics needs and priorities to the RAC. We need to clarify how these roles work together and what, if anything is missing.

Kaushik (and Armen) bring monitoring and reporting information needed by the RAC.

Discussion: How to make the RAC effective

relationship to CREM

There is some overlapping membership (Alexei). The CREM members represent functions not regions so there is no immediate way to seek a regional representation.

Do we need to identify a "core group"?

See below

Meeting time and frequency?

Full RAC meetings will be held monthly (Friday morning seems preferred) to address strategic issues and validate or adjust decisions of the RAC core group.

A core group of RAC members will be proposed by Richard after further discussions. The core group would comprise at least one person from each of physics requirements, facilities and operations and would meet weekly. All RAC members would be welcome at the weekly meetings, but the RAC core group would be required.

Kaushik raised the issue of urgent operational decisions (e.g. freeing up disk space). The agreed hierarchy is:

  • Operations people should refer operational resource-allocation issues to the RAC (core group), but are empowered to take any action they deem appropriate while waiting for the RAC response;
  • The RAC core group is empowered to make urgent allocation decisions, while reporting the issues and decisions to the full RAC for later endorsement or amendment.

Discussion: What is required for an effective resource allocation process?

Kaushik pointed out that monitoring and reporting on usage at an appropriate level of detail was essential to providing the RAC and US collaborators with the information needed to make decisions.

Richard's first draft of RAC activities follows with a draft of this addition:

  • physics needs:
    • prioritization of physics goals
    • express stream?? or similar?? (Currently ATLAS will not distribute the Express Stream)
    • characterization of the processes needed to meet physics goals (simulation, calibration, dataset production ...)
    • quantification of the computing needed for these processes
      • CPU
      • storage
      • bandwidth to storage
      • other

  • available resources
    • quantification of WLCG pledges and their likely evolution
    • quantification of allocatable resources
      • CPU
      • storage
      • bandwidth to storage
      • other
    • quantification and characterization of resources that are not centrally managed (aka Tier 3)
      • desires or offers to provide some of the computing needed for US-ATLAS physics goals
      • advice on the appropriateness of offered resources

  • how are resources being used:
    • monitoring and reporting on
      • CPU usage at the physics goal/process level of detail in addition to site/user
      • storage usage
      • WAN and storage bandwidth usage

  • communication (update and maintain Wiki)
    • resource availability
    • allocation process
    • allocation decisions

It is urgent that responsibilities for providing these functions be mapped on to RAC members. Several comments during the meeting (e.g. from Michael) pointed to vital parts of the above list of activities that were not currently happening.

AOB

Torre noted that the RAC already had an action item to create and own a process that would generate a prioritized set of needed but non-urgent simulation tasks that were ready to run whenever resources were idle.

Action Items/Homework:

  1. All (but especially Kevin/Jim/Ian), Propose the parts of the activity list above that are your responsibility and identify any activities that do not appear to be an appropriate task for any of the current RAC members.
  2. Richard Propose RAC core group membership
  3. Richard Non-urgent simulation tasks - consult with physics groups on the best way to establish two-way communication with the RAC to meet this and other needs.

About This Site

Please note that this site is a content mirror of the BNL US ATLAS TWiki. To edit the content of this page, click the Edit this page button at the top of the page and log in with your US ATLAS computing account name and password.


Attachments

 
Powered by TWiki
This site is powered by the TWiki collaboration platformCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback