[Fwd: Re: [evlatests] Re: EVLA tests, Aug 22 2005]

[Rich Moeser]

Barry asked me to forward his reply to the list. See below.

-------- Original Message --------
Subject: Re: [evlatests] Re: EVLA tests, Aug 22 2005
Date: Mon, 22 Aug 2005 16:59:10 -0600 (MDT)
From: Barry Clark <bclark at aoc.nrao.edu>
To: rmoeser at nrao.edu

> 
> > 
> > When I realized that Rich was no longer working, I asked the
> > operator to abort the script they were running so that I could
> > do other work.  When I started a script to look at 3C273 I found
> > that the antenna was still getting commands to go somewhere else.
> > These extra commands only went away when I asked to operator
> > to kill the process in which Rich's executor was running.  It
> > appears that there are still problems in getting the whole thing 
> > to shut down cleanly.
> 
> That's right, we ran into the same problem this morning during operator 
> test time. When a user sends a request to "abort" a script the executor 
> servlet sends a "halt" message to Barry's executor. The executor is 
> written such that the script will not actually abort until the next 
> source change which could be seconds, minutes or hours away. Until this 
> is corrected it will continue to be a problem. Also, as Pat mentioned in 
> her email, we found that the mib was generating acu commands even when 
> the file was aborted and the executor stopped.
> 
That has been, theoretically, fixed.  We need to have a closer look at
what's happening.  And yes, the MIB is supposed to do that for five minutes.