16 June 2008, Revision 9

Brief Status of Atlas Production Cache 14.1.0.Y ( candidate)

Andreu Pacheco-Pages / IFAE-CERN

Brief Status of Atlas Production Cache 14.1.0.Y ( candidate)

Current situation: (23 May 2008 10h37)

Cache is public since 9 June 2008. No date for next cache.

Proposed strategy: (23 May 2008 10h37)

Follow and create bugs in savannah for RTT, FCT and grid validation jobs.

NICOS tests Latest build: rel_3 (18 June 2008)

  • rel_1.Builds Ok. 57%
  • rel_2. Builds OK. 57%
  • rel_3. Builds OK.

RTT Tests (rel_3, 18 June 2008)

  • Information:
    • There is a mailing list for RTT issues: hn-atlas-runTimeTester@cern.ch
  • EvgenJobTransforms errors:

    • BUG#36635:TRF_UNKNOWN, producer=csc_genAtlfast, who=PoolSvc,message=POOL commit failed 0x9a769b0.

      • Affects Atlfast.5870.ttH_poslepnu_jj_bb

      • Incorrect configuration.

      • Submitted on 16 May 2008 and Assigned to Simon Dean.
      • 17 Jun 2008: No progress.
      • 18 Jun 2008: Fails again.
  • RecJobTransform errors:
    • TOBEIGNORED: TRF_UNKNOWN, severity=ERROR, who=IOVDbSvc, message=getAddress> Could not find IOVPayloadContainer for folder /GLOBAL/DETSTATUS/LBSUMM at time [52271,4:4294967295000000000], producer=csc_recoAOD
    • TOBEIGNORED BUG#37513: TRF_UNKNOWN, severity=ERROR, who=AnalysisTagBuilder.TopPhysTagTool, message=Collection Cone4H1TowerJets not found in StoreGate, producer=csc_buildTAG
      • RecJobTransforms/RecoTagTransf_130003 @ rel_0
      • 17 Jun 2008: Pierre-Antoine Delsart thinks it is impossible to fix in 14.1.x versions : reading rel 13 AOD requires ParticleJet to Jet converter which is only in 14.2.0
  • SimuJobTransforms: All tests ok.

FCT Tests AtlasProduction

  • Pileup pcache (rel_3, finished 18 June)
    • All tests OK.
  • Basic-pcache (rel_2, finished 3 June 2008)

    • BUG#37424 (37356) Atlas Reconstruction (Atlas Validation): TRF_UNKNOWN, severity=ERROR, who=CBNT_AtlfastMuon, message=Could not retrieve MuonContainer :AtlfastMuonCollection, producer=csc_atlfast

      • Affects Atlfast-CSC.005200.T1_McAtNlo_Jimmy-pcache_14.1.0.Y.rel_4-12Jun08
      • Bug opened 5 June 2008. Assigned to Simon Dean.
      • 17 Jun 2008: No progress.
  • Long pcache (rel_3 finished 11 June)

    • All tests OK.

Open Issues (9 June 2008)

ISSUE 080516-1025: Many bugs related with memory allocation problems. There is a high incidence of bugs opened due to crashes after failing to allocate memory. Batch systems usually limit the virtual memory of jobs exceeding 2.2-2.4 GB. This causes malloc() to fail when asking for more memory and then Athena crashes trying to use the returned pointer which is of course invalid. This will always happen, so it would be worth to handle the malloc failures in a way that the error can be easily identified to improve bug reporting.

PROCEDURES TO BE FOLLOWED (14 May 2008, 14:51)

  1. Look at RTT tests called Event,Simu or Reco JobTransforms,open and follow up tickets.

  2. Look at FCT tests daily

  3. Recommended that tags go into AtlasPoint1 first and then in AtlasProduction with the exception of Simulation tags.

  4. Follow up bugs and check daily their status.

  5. On Tuesday before 12am fill the Atlas Software Validation report for Job Transforms.

  6. Ignacio Aracena will advise the acceptance of Trigger tags into Atlas Production since May 14th 2008

  7. In bug reports specify failure rate whenever possible

  8. Expert in Atlfast I - Simon Dean

  9. Expert in Atlfast II - Michael Duehrsenn

  10. Technical changes to FCT must be sent to atlas-project-fullchaintest-technical@cern.ch

-- AndresPacheco - 12 June 2008

Edit | Attach | Watch | Print version | History: r7 < r6 < r5 < r4 < r3 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r7 - 2008-06-18 - unknown
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2020 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback