• Functionality
    • Pre-WS (performed once per release cycle)
      • globus_gram_client_test
      • globus_gram_protocol_test
      • globus_gram_job_manager_test
        • Fork
          • host?
        • PBS
          • lucky0.mcs.anl.gov
        • Condor
          • lucky0.mcs.anl.gov
          • submit #7
          • submit #10-13
          • submit #29
          • submit #36-38
          • submit #40-41
          • submit #43
          • submit #46-49
          • failure #1 Failed
          • failure #3-4
          • failure #53
        • LSF
          • Enable email responses instead of printing stats on stdout/stderr, so that the file comparisons work. Add the following -N line to the lsf.pm script as shown: 
                            print JOB '#BSUB -n ', $description->count(), "\n";
                print JOB '#BSUB -N', "\n";
          • dc-user2.isi.edu
    • WS (Fork and non-scheduler-specific tests performed for each release candidate, otherwise only once per release cycle)
      • globus_wsrf_gram_scheduler_test
        • Fork
          • ruly.mcs.anl.gov
          • submit205 "Bad Substitution" Failed
          • submit208 "Relative directory test" Failed (globusrun-ws output: globusrun-ws: Job failed: Invalid directory path "tmp".)
        • PBS
          • lucky0.mcs.anl.gov
          • submit002 "False" Failed (exit code not propagated from scheduler)
          • submit003 "Check Exit Code" Failed (exit code not propagated from scheduler)
          • submit205 "Bad Substitution" Failed
          • submit208 "Relative directory test" Failed (globusrun-ws: Job failed: Invalid directory path "tmp".)
          • failure005-008 "Unwritable std*" Failed
        • Condor
          • lucky0.mcs.anl.gov
          • submit003 "Check Exit Code" Failed
          • submit004 "Test multiple arguments" Failed
          • submit201-205 "* substitution" Failed (Condor quoting issue)
          • submit206 "relative path to executable" Failed
          • submit208 "Relative directory test" Failed (globusrun-ws: Job failed: Invalid directory path "tmp".)
          • failure001 "Executable does not exist" Failed
          • failure003 "Directory does not exist" Failed
          • failure005-008 "Unwritable std*" Failed
          • failure009-010 "stdin *" Failed
        • LSF
          • dc-user2.isi.edu
          • submit205 "Bad Substitution" Failed
          • failure005-008 "Unwritable std*" Failed
      • globus_wsrf_gram_service_test_unit
      • globus_scheduler_event_generator_test
      • globus_scheduler_event_generator_fork_test
      • globus_scheduler_event_generator_pbs_test
      • globus_scheduler_event_generator_lsf_test
      • globus_scheduler_event_generator_condor_test
      • globus_gridmap_and_execute_test
      • globus_c_wsrf_rendezvous_test
      • Basic multi-job
      • Basic multi-user
        • single job
        • multi-job
      • Recoverability
        • Basic (non-staging)
        • Recoverability tester (with staging)
          • Hangs in staging states caused by StateMachine and RFT bugs.
          • Fixes committed to gram_recoverability_tester_branch and will be available in 4.0.1.
  • WS Performance (performed once per release cycle, usually after the release candidate)
    • globus_wsrf_gram_throughput_test
      • non-delegation job
        • RESULT SUMMARY: Throughput peaks at 74 jobs per minute (1.23 jobs per second) for 2 client threads, each running with a load of 8. The throughput starts out at about twice that for 3.9.5 with 1 client thread, but drops off and errors out (OutOfMemoryError on server) earlier.
      • shared delegation job
        • RESULT SUMMARY: With a reasonable margin of error, the results for shared delegation are identical to those of no delegation.
      • shared delegation and stage in, out, and clean up job
        • RESULT SUMMARY: Throughput peaks at 11 jobs per minute (0.183 jobs per second) for a range of cases. These results can not be compared with the partial results obtained for earlier releases since different numbers of staging operations were done.

bullet key
  • Test with sub-test
  • Test or sub-test without any sub-test
  • Details about the above test
color key
Test has not completed
Test completed successfully
Test completed with partial success
Test failed