Native JDK random # seed

[Jan Grant]

On Thu, 16 Sep 2004, Cameron C Thomson wrote:

> 
>    I'm running FreeBSD 4.10 with jdk 1.4.2p6.
>    Summary: The na= tive JDK should provide the same stack trace error as
>    the linux-sun-jdk whe= n it suffers from a lack of entropy, rather
>    than simply locking with n= o output.
>    I had a number of apps that locked solid with no debug out= put or
>    stack trace.  I ran these apps using the linux-sun-jdk and reci= eved
>    an error relating to seeding random numbers (I can recreate and paste
>    = the stack trace if needed).  I then switched back to the native jdk
>    an= d seeded /dev/random with 'rndcontrol -s 10' . Now the native JDK
>    gets past= the lockup.
>    P.S. My solution to this is to set rndcontro= l in a startup script --
>    let me know if I'm missing something..
>    Tha= nks,
>    - C. Colby Thomson
>    J.D. Edwards Design Studio
>    [1]cthomson at unlnotes.unl.edu= 

For what it's worth, this reminds me of a lack-of-entropy problem that I 
sorted out on -4-stable with the native jdk. In the end I turned on some 
more entropy harvesting using rndcontrol; but my problem with attempting 
to track down the issue was that attempts to use the java debugger to 
get a trace of the offending thread wedged solid.


-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 http://ioctl.org/jan/
Not as randy or clumsom as a blaster.