Skip to content

Example of an UPV week-long run with many "Error in <TFile::TFile>: file <some Omicron root file>.root does not exist"

In https://scientists.virgo-gw.eu/DataAnalysis/UPV/2024/20240530/last_week_1400450421-1401055221/upv.log there are a lot of error messages from two Omicron channels:

(...)
Error in <TFile::TFile>: file /data/online/omicron/V1/SPRB_B4_DC_OMICRON/14005/V1-SPRB_B4_DC_OMICRON-1400500172-308.root does not exist
(...) _<many lines skipped>_ (...)
Error in <TFile::TFile>: file /data/online/omicron/V1/SPRB_B4_DC_OMICRON/14005/V1-SPRB_B4_DC_OMICRON-1400599796-280.root does not exist
Error in <TFile::TFile>: file /data/online/omicron/V1/SPRB_B4_QD1_GALVO_H_CORR_OMICRON/14005/V1-SPRB_B4_QD1_GALVO_H_CORR_OMICRON-1400500060-308.root does not exist
(...) _<many lines skipped>_ (...)
Error in <TFile::TFile>: file /data/online/omicron/V1/SPRB_B4_QD1_GALVO_V_CORR_OMICRON/14005/V1-SPRB_B4_QD1_GALVO_V_CORR_OMICRON-1400599964-308.root does not exist
(...)

=> A few questions:

  • Could it be that those Omicron ROOT files were being migrated from /data/online/ to /data/prod/ at the time?

  • How does the temporary redundancy between online and prod work in such cases?

  • Could there be that the existence of the ROOT files is checked at time t_1 but that those files are only used at time t_2>t_1? Meaning that the files existing at t_1 may have disappeared in the meantime!?

  • Would it be safer / simpler to run upv daily on older data (100,000+ seconds + some safety margin, if I remember correctly), which are guaranteed to be already in /data/prod/?