On Thu, 16 Jan 2003, Matthew Gregan wrote:
...
> Processes that are hung in D state are unkillable without divine
> intervention, or a reboot. D is an uninterruptible wait, a process
> hung in this state can be caused by a kernel bug, a hardware
> failure, or otherwise being hung inside the kernel (think of NFS
> mounts).
...

Yes, that can be troublesome. Several years ago I had that problem
with the HP-PA-RISC machines running HP-UX: When they waited for
something to come in on a serial port being accessed by kermit, and
there was nothing coming in, the getty (or whatever it was called), or
maybe even some serial driver in the kernel, deadlocked. The symptom
was that the whole machine came to an absolute standstill, and the
only thing that brought it back was pushing the button behind the lid.
No, even sending appropriate data on the serial line later did not
help... Must have been 1992-93.

Cheers,

Helmut.

+----------------+
| Helmut Walle   |
| [EMAIL PROTECTED] |
+----------------+

Reply via email to