George Greer wrote on 2009-09-28: > On Sun, 27 Sep 2009, Nicholas Clark wrote: > >> On Sat, Sep 26, 2009 at 12:36:00PM -0400, George Greer wrote: >>> Automated smoke report for 5.11.0 patch >>> c9a34126d2ca805f06cca8adaf9d74d2bf4f43ac blead-2433-gc9a3412 >>> perl-win2k: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz(~2490 MHz) (x86/1 > cpu) >>> on MSWin32 - Win2000 SP4 >>> using cl version 14.00.50727.42 >>> smoketime 3 hours 32 minutes (average 26 minutes 31 seconds) >> >> On Sun, Sep 27, 2009 at 03:35:00PM -0400, George Greer wrote: >>> Automated smoke report for 5.11.0 patch >>> fe0fb48b9e1689ad6e74e0516047f20e411bc3b4 blead-2457-gfe0fb48 >>> perl-win2k: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz(~2490 MHz) (x86/1 > cpu) >>> on MSWin32 - Win2000 SP4 >>> using cl version 14.00.50727.42 >>> smoketime 19 hours 28 minutes (average 2 hours 26 minutes) >> >> I didn't like *that*, did it? :-( >> >> I wonder why. > It's probably nothing you did specifically. I'm running TEST_JOBS=3 and > sometimes the tests hang and I have to manually kill one (doesn't seem > to matter which) to get it moving again. It's about once every two > weeks or so. That's also why my tests aren't in alphabetical order in > the log. > > In this case it was during the "-Dusedevel -Duseithreads -DDEBUGGING" > run and I killed the "openpid" test. I should've written down what the > other two live tests were but I was in a hurry to leave. I'll try to > investigate more next time that I'm not in a rush. I had a lot of trouble in the past with tests hanging in my smokers, and Jerry Hedden kindly came up with a solution: the Test::watchdog() function. That function is called from several test scripts that are liable to hang (a couple in each of: IO, threads, and threads/shared), and should kill them if they hang around too long. Perhaps we need to add watchdog() to a few more test scripts?Thread Previous | Thread Next