Dmitry Kalinkin (8ef0ca93) at 27 Mar 22:47
.gitlab-ci.yml: dispatch Snakemake.8 by renaming the file
... and 1 more commit
But why only the /root
? Maybe "EIC_SHELL_PREFIX" should also go there?
Dmitry Kalinkin (176571f6) at 03 Mar 03:14
Dmitry Kalinkin (1377b2d7) at 03 Mar 03:14
full_cal_reco.py: drop tile readout field
Support for https://github.com/eic/epic/pull/588
Dmitry Kalinkin (176571f6) at 03 Mar 02:53
Update full_cal_reco.py
Support for https://github.com/eic/epic/pull/588
Dmitry Kalinkin (83b53e63) at 03 Mar 01:46
full_cal_reco.py: drop tile readout field
Dmitry Kalinkin (980d3491) at 21 Feb 17:04
fix: avoid reading unused outputTrackParameters columns
Dmitry Kalinkin (66f82892) at 15 Feb 19:40
I'm open to migrating to XRootD. What is stopping that is that the required background files are currently missing there, and we'd need to convert them to ROOT format. Ideally, we'd support both cases to allow users some flexibility.
Thanks for hotfixing this. I'm not sure how one could read from S3 directly without going for FUSE / s3fs. Perhaps, we could agree to store large files in /scratch/cache/...
.
The hcal_barrel_protons jobs is the longest in the pipelines here. It is due to a set of serial simulations that are not able to be parallelized by the gitlab scheduler. This moves them into more atomic jobs.
This should fix the failing ZDC runs.