Subject: BMON
Ioanni.Katsavounidis@lngs.infn.it
Date: Thu Dec 16 1999 - 11:23:51 EST
Ciao Sandro,
I'm working on porting the whole BMON code to the external labs.
That includes .FOR and .COM files. In doing so, I think I've found
some minor bugs in the file
$1$DIA2:[BMONITOR.EXECUT]BMON.COM
which have been fixed in
$1$DIA2:[BMONITOR.EXECUT]BMON_NEW.COM
We have been receiving multiple copies of the output files of BMON
without exactly understanding why. I found the problem to be in
the logic that checks the queue whether BMON is *already* executing.
Please, do the following test
(1) Connect to VXMACA::MACRORUN
(2) Go to $1$DIA2:[BMONITOR.EXECUT]
(3) Do: SUBMIT/NOPRINT/KEEP/NOLOG/QUEUE=VXMACA_BATCH/NAME=BMON DUMB.COM
(this submits a "wait 00:01:00" command to fake the presence of BMON
in the appropriate queue). You now have 1 minute to perform the
next 2 steps !
(4) DO: @BMON_OLD_TEST -> this is effectively the standard BMON.COM file
with the RUN 'exec_dir'bmonitor line commented out, so that no real
processing is taking place. It also has commented out the DELETE
commands for intermediate files (needed for debugging.)
You will notice that the .COM file proceeds, ignoring the presence
of a BMON job on VXMACA_BATCH, as it should...
(5) Then, do: @BMON_NEW_TEST -> This does the same as the above, but
using the new version of the script I've modified. You should
get a message "JOB_FOUND=YES", and thus the code doesn't proceed.
(6) If you do the same comparison between BMON_OLD_TEST and BMON_NEW_TEST
when no BMON is running, you will see that they produce identical
files (except that the _NEW_ version produces 3 and the _OLD_ 4)
(7) Once you've convinced yourself that the _NEW_ script is good, just
rename the file BMON_NEW.COM to BMON.COM and it should immediately
take effect.
(8) Remember to cancel all temp files once you are done.
Let me know if there are any problems or if you have questions.
A presto,
Ioannis
This archive was generated by hypermail 2a24 : Thu Dec 16 1999 - 11:24:02 EST