cmssw: Strange output when firstLuminosityBlockForEachRun is used for Run-Dependent MC

Looking on last relvals of Run-Dependent MC, https://dmytro.web.cern.ch/dmytro/cmsprodmon/workflows.php?prep_id=CMSSW_11_1_0_pre5__RD_Harvesting_6HS-1586551827-ZEE_13UP18_RD

There is a strange in RUN number shows up in output, not sure if I understand how overwriten on run number should work. RelVal was submitted with following setting in DIGI step, process.source.firstLuminosityBlockForEachRun = cms.untracked.VLuminosityBlockID(*[cms.LuminosityBlockID(x,y) for x,y in ((315257, 1), (316082, 222), (316720, 445), (317527, 668), (320917, 890), (321414, 1112), (321973, 1334), (322492, 1556), (324245, 1779))])

What RUN number should be for input (GEN-SIM) with Run-1, LS 378? Should it be 316082? When I look on MiniAOD, DAS Run = 315257 was assigned instead of 316082.

Do I misunderstand how the code should work? Thanks in advance for the clarification.

Thanks to @schneiml for a first look on harvesting and spot this issue out from harvesting step.

FYI @jfernan2 @fioriNTU @christopheralanwest @ameyer-cern @tlampen @tocheng @chayanit @pgunnell @wajidalikhan