Re: unkillable ps processes

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Thu, 15 May 2025 09:45:00 +0200
    From:        Edgar Fuß <ef%math.uni-bonn.de@localhost>
    Message-ID:  <aCWbfFS47EPpwCtX%trav.math.uni-bonn.de@localhost>

  | I have (on -10) processes (running ps, in case that matters) that are running, 
  | consume 99% CPU time and are un-kill-able.
  | ktruss shows no system calls, attaching gdb hangs.

That almost guarantees that the processes are in some spin loop
inside the kernel (already attempting to do some system call).

A likely cause would be waiting upon some spin lock which isn't
being released, another might be a loop in some data struct that
is being searched.

Try using crash(8) and get a backtrace of where those processes
are, perhaps several times for at least one of them (unless the
backtraces are all the same) - that should provide some kind of
clue as to where in the kernel they're looping - the point of
several backtraces is to avoid concentrating on one function where
the process seems to be, if the actual loop is somewhere else
and the detected function is just being called a lot.

kre