[LU-18605] sanityn/31 fails with "No such file or directory"

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.17.0
• Severity: 3
• Assignee: Alex Zhuravlev
• Reporter: Alex Zhuravlev

Description

sometimes sanityn/31 fails the following way:

== sanityn test 31s: open should not revalidate invalid dentry ========================================================== 13:05:26 (1735563926)
sanityn.sh: line 1218: cd: /mnt/lustre/d31s.sanityn: No such file or directory
tail: cannot open 'f31s.sanityn' for reading: No such file or directory
tail: no files remaining
 sanityn test_31s: @@@@@@ FAIL: fail to stop 21777 
  Trace dump:
  = ./../tests/test-framework.sh:7229:error()
  = sanityn.sh:1226:test_31s()
  = ./../tests/test-framework.sh:7602:run_one()
  = ./../tests/test-framework.sh:7665:run_one_logged()
  = ./../tests/test-framework.sh:7483:run_test()
  = sanityn.sh:1233:main()

this is due to a race:

        ( cd $DIR/$tdir; tail -f $tfile || $MULTIOP . Dc) & pid=$!

        stack_trap "pkill -P $pid 2> /dev/null" ERR

        mv $DIR2/$tdir $DIR2/$tdir.old || error "mv $tdir $tdir.old failed"
        mkdir_on_mdt0 $DIR2/$tdir || error "mkdir $DIR2/$tdir failed"
        mv $DIR2/$tdir.old $DIR2/$tdir/ || error "mv $tdir.old $tdir/ failed"

where mv wins and starts before preceeding tail -f $tfile

Issue Links

• is duplicated by: LU-18297 sanityn test_31s: fail to stop