Hi!
I'm having trouble with my Linux box. When I boot the machine,
it starts fine, runs fine for about 1 hour to 2 days, then some
processes hang - ps shows status D. About at the same time,
every qmail-queue process that starts up ends up as zombie.
I did a ps -auxwl and it shows that the zombie processes are
in do_exit kernel function and the D functions are either in
do_done or wait_for_super.
Any idea? Maybe my hardware is faulty, but I've tried to change
the motherboard and the processor with no result. I've not yet
had a chance to change the memory - could it be faulty SIMMs?
This thing did not happen until a week ago. I've tried kernels
2.0.27 and recently upgraded to 2.0.30.
Please respond, I'm desperate about it. It's an ISP server and
the clients are unhappy :(
I have seen similar behaviour on my sparclinux box, some processes
just get caught in disksleep, and their children become zombies. It
seems to have to do with any heavy file access especially over NFS,
but not limited to it of course.
Since I am seeing it on sparclinux it seems like a generic kernel
problem. Just out of curiosity is the box you have SMP or UP, and are
you running an SMP kernel? My machine is a Dual processor hyper sparc
running an SMP kernel.
-- Manish Vachharajani <mvachhar@noc.rutgers.edu> Rutgers University http://www-no.rutgers.edu/~mvachhar Telecommunications Division Systems Programmer