On Fri, 10 Feb 2012 14:30:03 +0000, Charles Jardine <c...@cam.ac.uk> wrote:
> On 10/02/12 13:32, H.Merijn Brand wrote: > > Preparing a new database machine ... > > > > Do I need to worry? > > t/10general.t ............ 1/30 > > # Failed test 'system exit 1 should return 256' > > # at t/10general.t line 41. > > # got: '-1' > > # expected: '256' > > > > # Failed test 'system exit 0 should return 0' > > # at t/10general.t line 42. > > # got: '-1' > > # expected: '0' > > t/10general.t ............ 3/30 # Looks like you failed 2 tests of 30. > > t/10general.t ............ Dubious, test returned 2 (wstat 512, 0x200) > > This symptom indicates that the system built-in function is not working. > > Try > > perl -e 'print system("exit 1;"), "\n"' $ perl -e 'print system("exit 1;"), "\n"' 256 > If this reproduces the problem, you have something nothing to do > with databases to investigate. If it doesn't reproduce the problem, > it may be that Oracle is messing with the SIGCHLD signal. > > Are you connecting directly to the database using the bequeather? I've never heard of anything called a "bequeather" :) > If so, try connecting via SQL*Net instead. If avoiding the bequeather > fixes the problem, try putting 'bequeath_detach = yes' in your > sqlnet.ora file. This should allow you to use the bequeather and > system at the same time. above test was without TWO_TASK. Using the listener, I get All tests successful. Files=35, Tests=2008, 31 wallclock secs ( 0.40 usr 0.04 sys + 3.69 cusr 1.28 csys = 5.41 CPU) Result: PASS PERL_DL_NONLAZY=1 /pro/bin/perl "-Iblib/lib" "-Iblib/arch" test.pl -- H.Merijn Brand http://tux.nl Perl Monger http://amsterdam.pm.org/ using perl5.00307 .. 5.14 porting perl5 on HP-UX, AIX, and openSUSE http://mirrors.develooper.com/hpux/ http://www.test-smoke.org/ http://qa.perl.org http://www.goldmark.org/jeff/stupid-disclaimers/