[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: can't kill a PID



On Tue, 2002-11-05 at 01:28, Bob Proulx wrote:
> Matthew Gregan <kinetik@orcon.net.nz> [2002-11-04 18:58:41 +1300]:
> > > kosuke    9026  0.0  0.9 14460 4932 ? D  00:16   0:00 xmms
> > 
> > The state ``D'' means uninterruptible.
> 
> Any idea why?  Blocked waiting for I/O perhaps?  A DMA event that has
> not completed?  It is a kernel question, but under what circumstances
> would a process get stuck in that state?

Um, say a user level task is accessing some file on a NFS mount thats
mounted 'hard'. Then root goes and Stop's the NFS server running on
another machine. Now those tasks working on these files drop to a 'D'
state as they wait forever for the NFS mount to come back. Even if you
bring it back they're fudge. Forever and permanently 'D'.

I would like it if the kernel did a clean up of old 'D' processes that
have been hanging around a while using up PID number space. That would
be nicer.

Crispin

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: