2D sky subtraction results in over subtracted images

When using the weekly build w.2022.10 and running reduceExposure with 2D sky subtraction enabled:

reduceExposure.py /projects/HSC/PFS/Subaru --calib /projects/HSC/PFS/Subaru/CALIB-SuNSS --rerun=hassans/sunss/v46588_calibsunss_20220307b --doraise --id
visit=46588 arm=b^r -c measurePsf.oversampleFactor=11 measurePsf.version=Aug0121_v1 measurePsf.targetSize=21 doWriteCalexp=True doSubtractSky2d=True doMeasurePsf=True isr.doFlat=False

Results in over-subtracted lines (see , and ).

Note that some of those images indicate misalignment issues indicative of a poor detectormap. The above run makes use of the calib CALIB-SuNSS-2021-06-16. Using calib CALIB-PFI-20211220 requires a fix to DRP (see PIPE2D-993). Using a local fix and running the above against the PFI calib results in an improved image . In any case the main purpose of this ticket is to address the over subtraction, not the detectormap issues.