Re: Some failures in 0.8.0b1

From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Tue Dec 02 2008 - 14:42:50 PST

  • Next message: Paul H. Hargrove: "Re: LAM: Checkpoint is correct, BUT cannot restart with LAM+BLCR"
    You are correct, Neal.
    
    The ptrace test failure is an indication of a problem checkpointing a 
    ptraced process (e.g. one with gdb attached) and therefore not likely to 
    affect most "normal usage" of BLCR.  In fact, the test appears to mostly 
    work, except that the ptracer cannot detach when it should.
    
    -Paul
    
    Neal Becker wrote:
    > On Monday 01 December 2008, Paul H. Hargrove wrote:
    >   
    >> Thank you Neal (and sorry for calling you "Neil" before),
    >>
    >>   The fix to the prctl test will appear in 0.8.0_b2 (probably later
    >> tonight).
    >>   The ptrace test failure is still a mystery to me.  I know the fc8 and
    >> fc9 kernels replace the ptrace implementation from the kernel.org
    >> kernels with one written in terms of Red Hat's "utrace".  However, I
    >> cannot pin down what the difference is between the two that leads to the
    >> BLCR test failure.  It is possible that there is a utrace bug, but I am
    >> not holding my breath.
    >>
    >>     
    >
    > I'm guessing the ptrace test failure is not critical to my normal usage of 
    > blcr?
    >
    >   
    
    
    -- 
    Paul H. Hargrove                          PHHargrove_at_lbl_dot_gov
    Future Technologies Group
    HPC Research Department                   Tel: +1-510-495-2352
    Lawrence Berkeley National Laboratory     Fax: +1-510-486-6900
    

  • Next message: Paul H. Hargrove: "Re: LAM: Checkpoint is correct, BUT cannot restart with LAM+BLCR"