Summary
- Symptom
- When a cross disperser move is requested, the XDANGL keyword
starts changing and never stops; it cycles continuously
within a certain range of values.
- Problem
- The cross disperser motor is confused.
- Solution
-
- From the window manager's pulldown menu, select the
HIRES Control Menu option, then release on the
item labeled Stop Cross-disperser Oscillation.
- Alternatively, type the following command on
lehoula:
stop_cd_drift
- If this does not halt the cross-disperser, contact your
support astronomer.
- Symptom
- During a non-dark exposure, the color-coded box on the exposure control GUI
stays black indicating a closed shutter.
- Problem
- The shutter is opening but not enough to trigger the limit switch.
The EXPTIME keyword in the fits header will also be fractions of
a second rather than the desired exposure time. This may occur during
humid conditions or when the shutter is failing.
- Solution
-
- No solution needed, unless the exposure level implies that
the shutter really was closed. Shutter replacement may be necessary.
- Symptom
- The rotator will not initialize.
- Problem
- The rotator daemon needs to be restarted or it is running as another user.
- Solution
-
- Use the
ct
command to see what daemons
are running (see example output below) under the HIRES
server (background menu -> Login Windows -> HIRES SERVER xterm.
- Try executing
hires restart watch_hirot
- If that doesn't work kill the offending processes using
kill -9
and then run hires start watch_hirot
Example ct
output on the xterm on the HIRES server:
[21] hires1@lehoula: ct
============================================================================
Checking HIRES tasks...
MANDATORY daemons on LEHOULA (owner 'hireseng'; 6 tasks):
traffic, infoman, watch_ccd, watch_expo, dispatcher2, lickserv2
OBSERVER processes on LEHOULA (after a full startup; owner 'YOU'; 4 tasks):
write_image, obsr, ds9, ds9relay
============================================================================
...Checking LEHOULA for necessary daemon tasks ...
OWNER PID TIME PROCESS
hireseng 9362 19:10:24 0:05 /local/kroot/bin/watch_hirot dcs
hireseng 29932 Aug 03 6:44 lickserv2 -s hiccd -port 5686 -wname h_watch_ccd -lsname h_lickserv2 -
hireseng 29800 Aug 03 15:11 traffic
hires1 14514 19:31:15 0:00 dispatcher2 -s expo -n 0 -l dy
hireseng 9302 19:10:21 0:00 /bin/csh /local/kroot/bin/watch_hirot_monitor
hireseng 9330 19:10:24 0:00 /bin/csh /local/kroot/bin/watch_hirot_rotinit
hireseng 9332 19:10:24 0:00 /bin/csh /local/kroot/bin/watch_hirot_rotstby
hireseng 9334 19:10:24 0:00 /bin/csh /local/kroot/bin/watch_hirot_rothalt
hires1 14606 19:31:20 0:00 /local/kroot/bin/watch_expo
hireseng 28169 Jul 25 8:14 watch_ccd /local/kroot/data/music/info.hiccd/header_info hiccd hires e
hireseng 29833 Aug 03 0:01 infoman h_infom v=hiccd
hireseng 29846 Aug 03 0:31 watch_ccd /local/kroot/data/music/info.hiccd/header_info hiccd hires e
...Checking LEHOULA for observer tasks ...
OWNER PID TIME PROCESS
hireseng 15483 19:34:12 0:26 /local/kroot/bin/obsr -2l /local/k
hires1 14921 19:32:12 0:22 /local/kroot/bin/obsr -2l /local/k
hireseng 21712 Aug 14 12:22 /local/kroot/bin/obsr -2l /local/k
hires1 14886 19:32:11 0:00 write_image -s hiccd -wcname h_watch_ccd -lsname h_lickserv2 -winame h
hires1 14898 19:32:12 0:10 /local/kroot/bin/obsr -2lv /local/
hires1 14963 19:32:44 0:00 /local/ucolick/NewDS9/3.0.2/saods9/lib/ucolick/ds9relay
- Symptom
- The exposure meter is not responding to commands or error windows
are popping up (e.g. "Heartbeat expo. CTRL0CLK has stopped").
- Problem
- The exposure meter is in an unhappy state and needs to be restarted.
- Solution
-
- Attempt to restart the watch_expo daemon:
hires restart watch_expo
- If that fails with an error (e.g.
Error: Can't read Exposure meter dispatcher
status: mread select timeout.
), then do an "End of Night Shutdown" and then
restart all the GUIs with the "Start ALL HIRES GUIs" from the background menu.
- Symptom
- Neither the "Disarm" button, nor the "System Stop" button work.
- Problem
- The internal cal lamps are on.
- Solution
- Turn off the cal lamps, then try again.
- Symptom
- The two channels of the guider are very different (15k on
left, 9k on right)
- Problem
- The camera biases have been set to 0 instead of default
levels (515, 460) or so
- Solution
-
- In MAGIQ OA UI, open "Extended Camera Controls"
- Click "Load Defaults"
- Symptom
- 1. Nothing happens when you try to launch the HIRES GUI (this is the new GUI that replaced the old XHIRES GUI)
- 2. The HIRES GUI is missing after Starting up all HIRES GUIs
- Problem
- The service: hiresplus is not running.
- Solution
-
- From an xterm as hireseng on hiresserver, issue this command "$RELDIR/etc/init.d/hiresplus start"
- Bring up the HIRES_GUI via the pulldown menu: HIRES Control Menu -> Subcomponents -> Start HIRES_GUI
- Symptom
- You are seeing warnings about the cryolevers not being in the correct state
- Problem
- They are filling the reserve dewar for HIRES, and you can't fill the HIRES CCD dewar while this is happening
- Or, the sensor that reads the lever state is reading in error
- Solution
-
- Verify what state the software thinks that the cyolevers are in: from a hiresserver xterm: "show -s hiccd dwrxfrsw". If it is happy and ready for filling, it should read state=0
- Otherwise, call the summit and have someone check on the levers. They may need to be toggled back and forth.
- Have the summit initate a manual fill, and report the problem
- Symptom
- You see warnings about the HIRES autofill process NOT running
- It may also be preventing a commanded fill (via software) to occur
- Problem
- The keyword is not set correctly
- Solution
- From an xterm as hireseng at hiresserver, set the keyword to be enabled: "modify -s hiccd autofill=1"
- From an xterm as hireseng on hiresserver, issue this command "$RELDIR/etc/init.d/hiresplus start"
- Bring up the HIRES_GUI via the pulldown menu: HIRES Control Menu -> Subcomponents -> Start HIRES_GUI
- Symptom
- GUIs in VNC windows lost their top bars, so the GUIs cannot be
moved, resized, or rearranged. Right click in the background is not responsive.
- Problem
- The window manager (fvwm) is hung or has died.
- Solution 1
- Kill existing window manager (if still running)
log in to VNC server as the user
- Restart the window manager
log in to VNC server as the user
setenv DISPLAY kaalualu:[vnc port]
/usr/local/bin/fvwm -f $HOME/.vnc/fvwm2rc &
- Solution 2
If the steps above failed to solve the problem, kill and restart the VNC session.
- Symptom
- The HIRES GUI doesn't update after commands are set. For example,
sending open all covers is commanded, but the GUI does not show them open.
- Problem
- The hiresplus service has died or is frozen
- Solution
- Restart hiresplus service:
$RELDIR/etc/init.d/hiresplus restart
- Symptom
- Arc lines seem too faint, shifted, or have a uneven spatial line profile with a faint secondary peak.
- Problem
- The lamp selection stage may have shifted.
- Solution
- Home the lamp selection stage by modify -s hires lselcal=homed