Running the pipeline at the JCMT is slightly different from running the pipeline off-line in Hilo or at
another institution. The oracdr_scuba command should be used with no arguments. In this case the
data are arriving now file by file and flag files are written when the data files are completely written to
disk so that the -loop flag option should be used so pipeline will wait for new observation files to
appear. This is the default if the pipeline is run without any observation numbers or loop specifiers. In
order to restart the pipeline from a later observation, the combination of -from NN -loop flag is
required.
Additionally, the fastest machine should be used for the reduction with a local disk. Ask your support
scientist for more information on this – it is planned that the oracdr_scuba command will
automatically locate the correct data disk and inform you of the optimal data reduction
computer.