Display COMPLETE DOCUMENT Scroll Up Scroll DOWN MORE! TOP

How do I kill a paupsearch batch run?

On the helix systems, some cpu-intensive programs like paupsearch, framesearch, and mfold have been set up to run remotely on our high-speed parallel computer. This is invisible to users, so that you run these programs just like any other. For an interactive run of these programs, you can kill the program by typing 'Ctrl-C' (hold down the Control key and press 'c') just as with any Unix program.

For a batch run, however, the program is running on another computer, and it is not as easy to kill the program. During normal working hours, you can call up the helix staff (4-DCRT) and ask them to kill it for you. If you want to kill it yourself, you need to rlogin to churn, find the process, and kill it. Sample session: 
helix% mfold -batch
 
>> Setting up remote run on churn.nih.gov...

MFold predicts optimal and suboptimal secondary structures for an RNA
molecule using the most recent energy minimization method of Zuker.

 (Linear) MFOLD what sequence ? mzeorp1a.gb_pl

                  Begin (* 1 *) ?  
                End (*  1382 *) ?  
 
 What should I call the energy matrix output file (* mzeorp1a.mfold *) ? 

 ** mfold will run as a batch or at job.

 ** mfold was submitted using the command:
    "  atnow   "

       warning: commands will be executed using /bin/sh
job 908896860.a at Tue Oct 20 11:21:00 1998
In the above session, I started an mfold run in batch. Note that the program told me: 

>> Setting up remote run on churn.nih.gov...
which is a major clue that the program is running remotely on churn. I now realize I put in the wrong sequence and want to kill this run before it goes for days. 
helix% rlogin churn
IRIX Release 6.2 IP25 churn
Copyright 1987-1996 Silicon Graphics, Inc. All Rights Reserved.
Last login: Tue Oct 20 09:52:40 EDT 1998 by susanc@chisos.cit.nih.gov
Share II resource management; Copyright (C) 1989-1995 Softway Pty Ltd
Share login on /dev/ttyq19.
Tue Oct 20 11:21:48 EDT 1998

churn% ps -fu susanc
     UID   PID  PPID  C    STIME TTY     TIME CMD
  susanc 29736 29735 19 11:21:45 pts/19  0:01 -tcsh 
  susanc 29634   840  0 11:21:05 ?       0:00 sh 
  susanc 29635 29634  0 11:21:06 ?       0:00 /bin/csh -f ./mfold_29611_1 
  susanc 29761 29736  8 11:22:19 pts/19  0:00 ps -fu susanc 
  susanc 28429 28428  0 09:52:40 pts/17  0:01 -tcsh 
  susanc 29636 29635 80 11:21:06 ?       1:05 /gcg/gcgbin/execute/mfold -Init=mfold_29611_1.init 
  susanc 29637 29635  0 11:21:06 ?       0:00 Mail -s mfold susanc 
  susanc 19125 19124  0 14:13:27 pts/13  0:01 -tcsh 

churn% kill 29636
churn% ps -fu susanc
  susanc 18938 18937  0 14:05:30 pts/12  0:04 -tcsh 
  susanc 29736 29735  0 11:21:45 pts/19  0:01 -tcsh 
  susanc 29776 29736  7 11:23:19 pts/19  0:00 ps -fu susanc 
churn% exit
helix%
So, as above, I rlogin'd to churn, looked for all the processes running under my name (with 'ps -fu susanc'), found the one that says /gcg/gcgbin/execute/mfold, and killed that one's process id. Then I typed 'ps -fu susanc' again to make sure it was gone, and exited churn back to helix. The batch process will send you email saying that it was killed.