ak wrote:
> "Filip Larsen" <> schrieb im Newsbeitrag
> news:bp7md7$1r95$...
[...]
>>It is almost a definition that a process or thread that is blocked on
>>I/O do not use CPU. Conversely, if you see a live process or thread that
>>uses no CPU at all you can be pretty sure that it is because it is
>>blocked on I/O.
>
>
> I/O blocking while accessing harddisk? Its really strange!
Not necessarilly. There are several factors that could cause something
like that, from OS-level file locking to heavy disk activity, to
Java-level synchronization. It could be a Java or OS bug, but there is
nowhere near enough information for me to conclude that.
>>There is nothing strange about it. Notice that blocking
>>can include memory swapping done by the operating system, although this
>>rarely blocks a process for long on an otherwise working machine.
>
> Hmm, this mashine had really much memory - 1 GB I think.
Irrelevant. The question is not how much physical memory is installed,
but how much is available, the demands of other applications, and the
details of the OS' memory management subsystem. An OS might swap out
portions of an idle application's code or data irregardless of other
applications' current demand for memory.
John Bollinger