Test VEGAS spectral line modes with Breakthrough Listen software branch
Using scripts identical to those used for VPM regression testing (TGBT15A _915)
Mode 1 tested with L-Band, no motion
Modes 2 and 22 tested with X-Band, no motion
If time, will also test KFPA/RAMPS configuration and observe OH in G34.25+0.14
Details
Session begins at 10:30
Ray switches to BTL release
Switch complete at approximately 11:05
First attempt to launch astrid results in python path error: File "/opt/local/lib/python2.7/site-packages/wx/stc.py", line 10, in <module> import _stc ImportError: No module named _stc
Was able to fix by doing unset PYTHONPATH source /home/gbt/gbt.bash source /home/gbt/sparrow/sparrow.bash
This has never happened with Ryan's regular python environment. Unclear if this is due to version change.
Scan number reverted to 49 after version switch. First scan is 49, last scan is 78
vegas_status launches but first reports the following bash error: /home/sandboxes/rcreager/VEGAS/vegas_devel/src/vegas_hpc/vegas_spec-hpc.bash: line 23: syntax error near unexpected token `<<<' /home/sandboxes/rcreager/VEGAS/vegas_devel/src/vegas_hpc/vegas_spec-hpc.bash: line 23: `<<<<<<< Updated upstream'
Apparently caused by git conflicts in vegas_spec.bash
Ray resolved conflicts and vegas_status launches without issue
https://vegasdisplay.gb.nrao.edu/is not currently running. Unclear if this is due to version change or a power outage that occurred overnight.
This seems to have started updating again at some point. I was not watching it closely enough to note when.
Scheduling block finishes at 11:48
I realized that many of the mode 1 tests under TGBT15A _915 were done at X-Band. Will do all remaining tests at X-Band. Ray suggests checking one mode from each group, i.e. 1-3, 4-9, 10-19, and 20-29. Will check modes 1, 2, 22 (for comparison with previous regression tests) as well as 4 and 10. In interest of time, reducing number of on/off pairs to 5.
Operator switches to X-Band at 11:49
Mode 1 (X-Band) tests begin at 11:51
Using CheckMode1 _rmp scheduling block (modified for X-Band)
Switching to KFPA to test a RAMPS configuration and on/off scan.
KFPA test begins at 13:01
Using onoff_L11_test scheduling block
Configuration error: ERROR: The number of frequencies provided must be equal to or greater than the value of nwin; 10 < 64
nwin does not appear to be set in /home/astro-util/projects/16A353/ramps/vegas_config_ramps_adv_test.py. Using a ResetConfig to see if that helps.
ResetConfig seemed to clear the above error. Scheduling block continuing normally.
Peforming one on/off pair, 30 s duration
Scans #129/130
Scheduling block finishes at 13:09
Switching to L-Band for OH test scan
L-Band OH test begins at 13:13
Using 02 calib_OH_tracking scheduling block
Should be setting up for mode 15 on banks C, D, E, and F
Received an abort when trying to start scan. Seems to be at the slew stage but there are no messages from the Antenna.
Operator tried a manual move. No change.
VEGAS reports a few errors:
Vegas activate inhibited because of illegal value (this seems to be tied to the value of number_spectra, which is being set to 0)
Valon frequency 1500 MHz is not equal to that set (on banks A, C, D, E, and F)
For some reason number_spectra is being set to zero in the Coordinator, but number_spectra on bank Managers is correctly set to 1
Bank A has number_spectra = 0 but it is not configured for data taking by this script. Manually changing number_spectra back to 1 on bank A and skipping any further configuration seems to allow the scan to continue.
Scan finally begins at 13:41.
Performing one frequency switched scan, 300 s duration
Scan #131
Scheduling block finishes at 13:46
Ray wants to try putting the system in BTL mode and then switching back to VEGAS to see if there are any issues going back and forth.
Trying to run an L-Band script after Ray put the system in BTL mode
Using CheckMode1 _rmp scheduling block (modified for L-Band)
First attempt resulted in an abort due to conflict between length of deltafreq and restfreq.
Added a ResetConfig and VEGAS configured properly.
Scan aborted on first attempt due to 'VEGAS HPC program taking too long to be ready'
Ray attempted to manually conform params and prepare but second attempt had same timeout error (removed ResetConfig for this attempt).
This somehow resolved itself. Ray says possibly an NFS issue.
Performing one on/off pair, 60 s each
Scans finally begin at 14:06
Scan #134
Sequence aborted at 14:07 to try BTL/VEGAS switch again.
Submitted SB again at 14:07 to test mode switch.
Scan #135
Second attempt to switch from BTL to VEGAS seems to have worked without issue. Note that no ResetConfig was used this time.
Scan aborted at 14:09
Ray began switch to release branch at 14:10. Completed at 14:26.
Conclusions
A by-eye inspection shows no issues with data quality.
Outstanding issues:
Astrid failed to launch under Ryan's default environment.
02 calib_OH_tracking resulted in number_spectra=0 bank A.
VEGAS HPC program failed to start on first attempt to switch between BTL and VEGAS modes.
Note that it is not clear if issue 1 and 2 are related to BTL branch.