STARTUP INSTRUCTIONS FOR RTS-F AND ROBOT Created 07/01/11 EF Updated 02/03/20 EF Type Rstatus in any xterm to find out whether RTS2-F is running. The RTS2-F servers should run all the time. Steps 1 and 2 are only necessary if you find RTS2-F is NOT running, or it becomes unresponsive during a run started with step 5 below. Type Rstatus in any xterm to find out whether RTS2-F is running. 1) If RTS2-F is stuck, make sure the robot is not running in Telshell (use CTRL-C if needed), then in any xterm type Rstop. NOTICE: Rstop clears the queues but it also saves them to a file that Rstart uses to reload the queues. Rstop may also be issued to stop Rob when manual observations are conducted. In this case, the cron jobs that monitor Rob are also stopped, and will need to be restarted. The crontab entries are self-explanatory. Use crontab -e to set up jobs, and crontab -l to list them. The example that follows shows the crontab with all the jobs enabled: -------------- >crontab -l # next 6 commented out after stopping Rob with Rstop # uncomment after Rstart 00 14 * * * /home/observer/RTS2-F/rts2.harvestcat > /dev/null 2>&1 05 14 * * * /home/observer/RTS2-F/rts2.gocircular >> /tmp/gocircular.log 2>&1 00 10-18/2 * * * /home/observer/RTS2-F/rts2.autosavequeue > /dev/null 2>&1 */10 * * * * /home/observer/RTS2-F/rts2.goheartbeat > /dev/null 2>&1 */5 * * * * /home/observer/RTS2-F/rts2.gocheck_httpd > /dev/null 2>&1 */1 * * * * /home/observer/RTS2-F/rts2.gogetALTAZ > /dev/null 2>&1 00 07 * * * /home/observer/M1temps/RunTempsAuto > /dev/null 2>&1 00 08 * * * /usr/local/bin/rms_me > /dev/null 2>&1 # greenhouse lights out starting 16/11/01 30 08 * * * /home/observer/Sky/RunLights > /dev/null 2>&1 ------------- Use crontab -e with extreme care!!! 2) To restart RTS2-F and the monitoring, first ensure that the real-time system is up, otherwise in any xterm type gokep; once the system is up, in any xterm type Rstart. Then type GOrobot in Telshell. Also type GOweather and GOqueues. Then click the box at the top of the GOweather GUI, to set the state to ON (or Standby if the weather is inclement). NOTICE: a) Rstart will reload the queues from the last file that Rstop saved. Double-check their contents. b) GOweather is different from STARTweather, which restarts weather sensing, see the next paragraph. 3) If the robot senses the weather is inclement but you are CERTAIN (using the MMT skycam and confirming with the 1.5m observer) that the weather is clear and the sensing is broken, you may use the command KILLweather to disable ALL weather sensing. Use with great caution. To restart weather sensing after a KILLweather or after any failure, type STARTweather. 4) To make sure it's all working, type MONrobot. This command drives an xterm and shows the status of the whole system. Navigate with the arrow keys on the left column (5 items should be visible); tab to the right panel, where you can change states where you see a W to the left of a variable; use the F9 key to see and interact with menu items at the top: e.g., to change the mode of the system (ON/STANDBY/OFF) under States. Use F10 to quit (CTRL-C if that fails). 2 tabs from the right panel get you back to the left column. Try man rts2-mon (MONrobot is aliased to that) for a complete description. 5) In the Telshell window once Rstatus shows RTS2-F is running, type GOrobot. That will start the main robot script. For safety, the default mode of the system is OFF, so it must be switched to ON to start the nightly sequence, preferably with the GOweather GUI. GOrobot available flags are: -D Skip dome flats -d Use alternative dome flats list: /home/observer/RTS2-F/FLATS/ALTFILT -e Skip evening sky flats -m Skip morning sky flats -R FORCE tele reference -F Skip focus run at the start -P Run FIXpoint at the start, to fix the pointing -E get 2 defocused images for collimation 6) On flwo48, watchdog jobs monitor the robot's activity. They verify that the robot is running, that data are being acquired at night, that the dome is not open under weather conditions over closing thresholds (wind, RH, precipitation, clouds) and that the CCD is operating normally. If there is any violation, the robot warns the 1.5m observer, and requests a clearing of the violation every 10 minutes, until the observer clears the violation and issues a clear_robot command on flwo48. On flwo60, /var/log/chk.robot.log shows the status of any violation every 10 minutes, and on flwo48, /var/log/rob.clear shows when the violation was cleared.