[vox-tech] unkillable processes (was: ac97 sound problems)

[vox-tech@lists.lugod.org]

On Thu, Apr 25, 2002 at 03:10:28PM -0700, Henry House wrote:
> On Thu, Apr 25, 2002 at 11:05:43AM -0700, Peter Jay Salzman wrote:
> [...]
> > as you point out, processes in "uninterruptable sleep" can't be killed
> > with SIGKILL.  the process is put to sleep while the kernel waits for
> > some event to happen.  this corresponds to process status "D".
> > 
> > as you point out, it can be kernel bug.  often a race condition.
> > but it can also be caused by hardware failure.
> 
> I have seen such unkillable processes crop up on NFS clients, in cases of
> server failure.

Henry,

  Very true... a valid example of an unkillable process (which is not
a kernel bug and is not a zombie ;).

Any other examples?


  That is related to NFS hard mounts.  (The process will wake as soon
as the NFS server becomes reachable and completes the pending request,
if you kill the process while it is waiting, the process will die the
second whatever write operation it was doing completes).


cut and pasted from part of a off channel thread:
=======
  The _only_ exception of where a long uninterruptable sleep is acceptable
is for NFS mounted files when the machine has lost access to it's file
server, this is because NFS by default will retry indefinitely for
the server to return, it is possible to "soft or intr" mount with to
disable this feature... even hard mounts can be force unmounted by root
which should wake all affected processes.