r2 - 10 Mar 2009 - 18:01:51 - KaushikDeYou are here: TWiki >  Admins Web > MinutesDataManageMar10

MinutesDataManageMar10

Introduction

Minutes of the US ATLAS Data Management meeting, Mar 10, 2009
  • Previous meetings and background : IntegrationProgram
  • Coordinates: Tuesdays, 3:00pm Central
    • (309) 946-5300, Access code: 735188; Dial *6 to mute/un-mute.

Attending

  • Meeting attendees: Bob, Wei, John, Wensheng, Armen, Hiro, Pedro, Charles, Patrick, Shawn
  • Apologies: Alexei
  • Guests: None

Topics for this week

  • BNL new storage status - Pedro
    • 5 Thorr's ready, each ~30 TB usable
    • Current performance is not very good - test underway
    • ConditionsDB? - provision 2 TB times 3 copies for next 6 months
    • Storage on ACAS nodes will go away by end of 2009
    • ~101 TB will be retired March 24th, almost all on BNLPANDA, most files are seldom accessed, so decide to simply retire them
  • Recent file corruptions - Shawn, Hiro
    • Large number of files from (at least) one compute node at AGLT2 had bad adler32
    • Jobs were finished successfully, so corrupted files are now in system
    • Shawn will prepare list of corrupted files, so they can be cleaned up
    • Software changes needed to prevent future problems:
      • Need new Panda sitemover to also check adler32 from dcache, before registering in LFC (Paul's todo-list)
      • Need FTS to check adler32 after transfer (in workplan - Wei is checking?)
      • BNLPANDA SS plug-in to check adler32 from BNL dcache against original, and fail tranfer if they do not match (Hiro + consultation with ADC)
    • John is seeing some corruption errors with RAW files - will follow up
  • Hot issues
    • GROUPDISK filling up - Shawn
      • Could be dark files, possibly hundreds of failed copies per file (random check of one file shows 114 copies)
      • ToA? seems to be wrong - causing all attempts to fail (grpdisk vs GROUPDISK)
      • Continue to follow-up: why no error seen in dashboard, why DQ2 trying so many times...
  • AOB
    • Change to earlier time slot - request from Alexei
      • Noon CDT work for everyone - will switch time starting next week


-- KaushikDe - 10 Mar 2009

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