Larry,
 
We have had this problem also, and it seems that a little command file
should do the trick. This works okay, so long as the console isn't being
used for running programs (eg. CONMAN). If you're using terminal
windows, it shouldn't be a problem.
 
Call the command CONCMD (or something)
 
COMMENT concmd
setvar cmdflag,0
while cmdflag=0
   input cmd;wait=300;prompt="SYSTEM: "
   if len(cmd)<>0 then
      continue
      !cmd
      setvar cmd,""
   endif
endwhile
 
This just loops an input over and over again on the console. This has
tested okay, only problem being that you get a quite a few "SYSTEM :"
prompts on the console (which isn't really a problem)
 
Best of luck
 
DAMO!
 
mailto:[log in to unmask]
 
 
 
 
Larry Akers wrote:
>
> Hello everyone,
>
>     We have been experiencing a problem about once every month or two where
> an operator enters an incomplete console command.  This can be anything from
> accidently hitting the space bar or any key and then not hitting the return
> key to just not finishing a command with return.  We have installed a patch
> on all of our systems (we have about 70 HP3ks in our shop) which keep the
> systems from crashing when the console buffer fills up.  However, we do not
> currently have any way to notify the operators that the console buffer is
> full.  This causes the system performance to degrade to the point where a
> job which normally completes in 1 minute can run for hours.  We are using
> OpC to monitor our systems and have many redundant checks built into the
> monitoring process, but we have not been able to come up with a way to
> notify the operators of this problem.  We are currently using 5 HP ux
> workstations to monitor about 70 HP3k's and about 100 HP9k's using OpC.
>
>     So my question is, does anyone know of any way to send console buffer
> full messages directly to OpC, or close the current console buffer when it
> reaches a certain limit, say 75%, and issue a console message that could be
> trapped by OpC and generate an alarm.  Sorry for the looong sentence.  This
> problem seems to always happen during critical customer times.
>
>     I suppose that there are many solutions, such as hiring a temp employee
> to go to each console every 10 - 15 minutes and hit return.  But with so
> many systems to monitor it seems that there should be an automated way to
> handle this problem.  We could open and iconize all of the console windows
> and then open another window for the operators to use for system access.
> But then our method of console log access (currently kept on an HP9k) would
> not work.
>
>     Anyway, this post has gotten very long so I'll end now.  If anyone has
> any suggestions that may help I would appreciate the advise.
>
>                                               REgards,,,LAkers...