[Clas12_calcom] [EXTERNAL] Removing unnecessary banks from the calibration schema

[Nathan Baltzell]

Note, this proposed 2x reduction of (unused?) calibration data would increase our lifetime on /volatile by 2x and greatly simplify disk logistics.


On Jan 13, 2023, at 12:47 PM, Raffaella De Vita <Raffaella.Devita at ge.infn.it<mailto:Raffaella.Devita at ge.infn.it>> wrote:

Dear All,

During recent data processing for calibration purposes, we have noticed frequent train jobs' failures due to the input data volume exceeding the requested disk space. The reconstruction output size with calibration schema is indeed quite large, which is not only causing these job failures but is also leading to large disk space usage on /volatile.

Checking the calibration output from a RG-A Spring19 run, we could reduce the file size by 40% dropping the following banks that do not seem to be used by calibration software:

Total file size: 6383 MB
- RAW::VTP          1108 MB 17%
- RECHB::* 443 MB  7%
- TimeBasedTrkg::TBCrosses      129 MB 2%
- TimeBasedTrkg::TBSegments     367 MB 6%
- TimeBasedTrkg::Trajectory     574 MB 9%
Total reduction 41%

Any objection?

Thanks,
Raffaella

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/clas12_calcom/attachments/20230114/7ee7b82f/attachment.html>