Needle in the haystack: same Java code on two identical machines, one passes one fails

[Osipov, Michael]

> Hi,
> 
> > On 16 Mar 2017, at 09:56, Osipov, Michael <michael.osipov at siemens.com>
> wrote:
> > […]
> >
> > Updated both machines:
> > FreeBSD blnn719x.ww004.siemens.net 10.3-STABLE FreeBSD 10.3-STABLE #2
> r315312: Wed Mar 15 19:59:58 CET 2017
> root at blnn719x.ww004.siemens.net:/usr/obj/usr/src/sys/BLNN719X  i386
> >
> > FreeBSD blnn714x.ww004.siemens.net 10.3-STABLE FreeBSD 10.3-STABLE #2
> r315312: Wed Mar 15 19:55:39 CET 2017
> root at blnn714x.ww004.siemens.net:/usr/obj/usr/src/sys/BLNN714X  i386
> >
> > No avail. It was worth a try, at least.
> >
> 
> Indeed. Given the symptoms, next thing I would do would be to check
> exactly what versions of shared libraries the java binaries on both sides
> are loading and chase down any differences.
> 
> If that doesn’t help, check that the CLASSPATHs are identical, check the
> versions of everything loaded from standard extensions or the CLASSPATH,
> and chase down any differences.

I actually did that yesterday. Compared MANIFEST.MF, Surefire start config
with clasapath, even used truss and compared. No difference apparently.

I also did prepend "stdbuf -i 0 -o 0" to all processes, no avail.

This one is really tricky. I won't reinstall the entire machine, I will,
Unfortunately, mark this machine/OS installation as defective.

Michael