java execution stalls indefinitely

[Kúti Zsolt]

Thu, 7 Jan 2010 11:18:13 +0200 -n
Achilleas Mantzios <achill at matrix.gatewaynet.com> írta:

> You could run both instances in the two machines with ktrace (1) or
> truss(1) to see what is the good instance running where the bad
> instance stalls.

Hello Achilleas,

Thanks for the hint! Running truss shows after some time a repeating
sequence of the following:

...
gettimeofday({1262867164.433795 },0x0)
= 0 (0x0) gettimeofday({1262867164.433851 },0x0)		 = 0
(0x0) clock_gettime(0,{1262867164.433889494 })	 = 0 (0x0)
_umtx_op(0x2820e260,0x8,0x1,0x2820e240,0xbf5e4cec,0x805b528) ERR#60
'Operation timed out'
...

Unfortunately I do not know what this all mean, what setting (if any)
can cause this behaviour.
I searched for _umtx_op on the net and found a similarly behaving python
and OOo case. There was no conclusion, just some suspect of some
threading problem.

I use some serious piece of java programs that work fine.
It is possible that this is a generic OS (ports?) problem and has
nothing to do with Java itself.


Thanks!
Zsolt

Ps.: The only notable difference I can see between the two boxes, that
the problem happens on a Gigabyte mother board with Intel(R) Core(TM)2
Duo CPU E8400, while the other is an Asus with AMD.