We have been having problems with defunct processes under OSF 3.0; these
are network programs which use a ptypair to remote boot programs to
lantronix boxes, which sometimes leave defuncts when they die. These
defuncts have the following odd properties:
They _are_ being waited for (their parent is specifically calling waitpid()
on each of them, in fact)
The tty itself is locked up; attempting to open the tty leaves a signal-immune
process (waiting in open, of course). Attempting to open the pty generates
an I/O error.
This is only an intermittent problem (we have many such processes, and
only a few do this, and not always the same ones), but it can be a major
problem. Does this sound familiar to anyone out there?
Received on Thu Sep 07 1995 - 19:26:07 NZST