wfdisp Displays

wfdisp Displays

———- Forwarded message ———- 
Date: Thu, 11 Sep 2008 10:43:34 -1000 (HST) 
From: Jonathan Kemp <j.kemp@jach.hawaii.edu>
To: Brad Cavanagh <b.cavanagh@jach.hawaii.edu>
    Tom Kerr <t.kerr@jach.hawaii.edu>
Cc: Andy Adamson <a.adamson@jach.hawaii.edu>
    Craig Walther <c.walther@jach.hawaii.edu>
Subject: wfdisp displays 

Hi guys, 

I finished configuring a new script this week to enable easy launching of the Gaia displays in quadruplicate on the new display server.  A brief note will go out to affected staff (ss/to), but here are some more details, for the record, and in case you are interested (perhaps only Brad is interested, and maybe not even him ðŸ˜‰ . 

Anyway, a few things had to be accomplished with the new display server setup: 

 – incorporation of .xinitrc_wfcam.csh which ran on wfdr[n] consoles with 
   its lightweight X server to a script so that $RTD_REMOTE_DIR and $HOME 
   are set correctly before starting Gaia; 

 – resolution of unique authentication issues related to moving this to a 
   remote display setup that is not run on the individual consoles; 

 – make this all happen in quadruplicate at the push of a button for ease 
   of use by observers who may not be familiar with resetting our new 
   display system. 

I wrote a script called wfcamGaia which can be run on wfdisp to stop/start Gaia in quadruplicate, or which can be run individually after connecting to a wfdr[n] machine in a terminal on wfdisp to instead set it up for a single camera. 

This script incorporates all the $RTD_REMOTE_DIR and $HOME configuration issues specific to each wfdr[n] machine so that they are each set appropriately to /ukirtdata/reduced/wfcam[n] as they were previously. This allows the ORAC-DR pipelines to find and use the four separate Gaia displays appropriately during nighttime reductions. 

I had to make a set of four symbolic links equivalent to /ukirtdata/reduced/wfcam[n]/.Xauthority -> /home/observer/.Xauthority to compensate for the reset $HOME variable now that remote connections are used to initiate Gaia instances that are displayed back to wfdisp. 

This is now incorporated in a single button/icon (in the style of ukirtObs/wfcamObs/cassObs, including appropriate machine name in parentheses) on the observer account task launcher panel that kills any existing ORAC-DR Gaias on wfdr[n] machines and restarts them.  It puts the four new Gaias in their appropriate places on wfdisp so that the arrangement is as it should be for the four cameras.  wfcamDisplay can still control the displays as it did previously.  And, the button/icon will not work if accidentally run from a console other than wfdisp or if accidentally run somehow from an account other than the observer account. 

So, some remaining issues/gotchas/to-watch-fors… 

– Because ORAC-DR now routinely uses two Gaias to enable group co-add 
  displays, the second Gaia instances for each wfdr[n] machine will appear 
  on wfdisp later after a wfcamGaia command is issued.  I am not certain 
  that they will show up in the appropriate places and they may have to be 
  rearranged by the observer if it is one of those infrequent instances of 
  needed to cycle the four ORAC-DR WFCAM Gaias.  However, I’m hoping that 
  the geometry of the previous Gaia will be inherited for each group Gaia 
  that pops up.  Just a warning for the first night’s observer. 

– I am doing remote Gaia display over SSH since that seemed simplest with 
  SSH taking care of all the display routing.  Test reductions at the 
  console showed that this seemed to keep up, but this is just a heads-up 
  as to how the bits are being sent from wfdr[n] machines back to wfdisp. 

– I did not incorporate the xset variables used by the wfdr[n] lightweight 
  X servers that disabled the screensaver and monitor Energy Star features. 

I think that’s it.  For now, the script also stops any existing Gaias just before launching new ones so as not to allow Gaia confusion or Gaia swamping. I *may* even set it to kill active pipelines, but I’m not yet sure about this. I have seen observers restart Gaias before without stopping pipelines only to find out all the Gaias start popping up on kauwa which wasn’t what they intended. 

Certainly let me know if you have any questions.  I’m around town for a few more days. 

Cheers, 

jonathan