Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-4818

Add debugging of bad read file descriptor when files are missing during recovery

    • Type: Icon: Bug Bug
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • WT3.2.1, 4.3.1, 4.2.0-rc3
    • Affects Version/s: None
    • Component/s: None
    • None
    • 3
    • Storage Engines 2019-07-19

      random_directio failure with core dumped.

      The spinlock test failed with a core dump generated, on kodkod-aws.

      http://build.wiredtiger.com:8080/job/wiredtiger-test-spinlock/4763/console

      CONFIG: ./test_random_directio -t 5 -m none -T 5
      Parent: Create 5 threads; sleep 5 seconds
      Create 5 writer threads
      Spawned 5 writer threads
      CONFIG: ./test_random_directio -t 5 -m none -T 5
      Parent: Create 5 threads; sleep 5 seconds
      Beginning cycle 1/5
      Suspend child
      Check DB
      Copy database home directory using direct I/O to run recovery,
      along with a saved 'pre-recovery' copy.
      Open database, run recovery and verify content
      starting full scan at 5000
      scanned to 9331
      
      (...)
      
      Beginning cycle 5/5
      Suspend child
      Check DB
      Copy database home directory using direct I/O to run recovery,
      along with a saved 'pre-recovery' copy.
      test_random_directio: FAILED: copy_directory/103: rfd >= 0
      process aborting
      Aborted (core dumped)
      FAIL random_directio/smoke.sh (exit status: 1) 

      The build_posix directory was archived on kodkod-aws with a filename "jenkins-wiredtiger-test-spinlock-4763.a610f847393f5d21923f92e38a7e1aa7258f7d47.tgz" under the job workspace.

            Assignee:
            sue.loverso@mongodb.com Susan LoVerso
            Reporter:
            luke.chen@mongodb.com Luke Chen
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated:
              Resolved: