Killing processes when "kill -9" doesn't work

David Yang d-yang at cs.columbia.edu
Wed May 2 14:00:19 AEST 1990


I was trying to run dbx and got impatient after waiting for about
5 min. for the initialization to finish.  I couldn't interrupt (ctrl-c) or
suspend the job (ctrl-z).  I then tried "kill" and "kill -9", but
the jobs remained.  I killed the shell process, but again nothing
happened to the dbx jobs.  I then foolishly tried dbx 3 more times
with the same result, so now I have 8 dbx-associated jobs sitting around.

1.  How can I get rid of these jobs?
2.  Is there a way to prevent this?
3.  Do the swapped out jobs affect the machine load, and if so, how much?

Below is the output of "ps ux":

USER       PID %CPU %MEM   SZ  RSS TT STAT START  TIME COMMAND
d-yang   21057 15.8  1.4  192  456 r1 R    23:57   0:00 ps ux
d-yang   11271  0.0  0.0   24    0 qc TW   Apr 30  0:00 b-splineinter
d-yang   20368  0.0  0.3   48   88 r1 I    23:47   0:00 sh /usr/local/bin/Pnews 
d-yang   20383  0.0  1.5  168  488 r1 S    23:47   0:01 /usr/ucb/vi /u/cs/d-yang
d-yang   19731  0.0  0.0   24    0 t1 TW   23:23   0:00 b-spline
d-yang   20367  0.0  0.3   32   88 r1 I    23:47   0:00 /bin/sh -c Pnews -h /u/c
d-yang   20058  0.0  0.0   56    8 r1 I    23:33   0:01 -ksh (ksh)
d-yang   19730  0.0  0.1  208   16 t1 I    23:23   0:00 dbx b-spline
d-yang   12577  0.0  0.0  208    0 qc IW   Apr 30  0:00 dbx binter
d-yang   12578  0.0  0.0   24    0 qc TW   Apr 30  0:00 binter
d-yang   11268  0.0  0.0  208    0 qc IW   Apr 30  0:00 dbx b-splineinter
d-yang    5831  0.0  0.0  208    0 sa IW   Apr 30  0:00 dbx b-splineinter
d-yang   20341  0.0  1.1  144  336 r1 I    23:46   0:00 rn
d-yang    5832  0.0  0.0   24    0 sa TW   Apr 30  0:00 b-splineinter


Any help/advice would be appreciated,
(e-mail would be preferred, especially if this is a worn-out topic--
 I didn't see it in the most recent "Commonly Asked Questions posting"
 I have-- sorry if I'm not reading carefully enough)
David Yang
d-yang at cs.columbia.edu



More information about the Comp.unix.questions mailing list