Segmentation violation when trying to remove GEC from SingleHighPTMuon HLT1 line (non-deterministic)
Report of a crash/segfault with Allen
Brief summary of encountered problem
Tried to add a new HLT1 line with no GEC and I see seg faults/failures/sometimes it works. Feels indeterministic/transient.
Dataset you were running on
Used a single input file (Z2mumu signal MC - roughly 5000 events).
Allen version used (or nightly, as appropriate)
Allen branch: "rjhunter-remove-gec-from-highptmuon". See !664 (closed)
Please include the platform Allen was compiled for (CPU or GPU as well as the type of build)
CPU default with lb-stack setup ("x86_64_v2-centos7-gcc11-opt")
Option file used to execute Allen
moore_options_file.py
terminal command: MooreAnalysis/run gaudirun.py moore_options_file.py
Additional information
I compiled Allen standalone to regenerate the sequence and then copied the sequence to replace hlt1_pp_default.json in the lb-stack. I then recompiled afterwards.
Steps after following instructions to compile Allen standalone:
git submodule update --init --recursive
git fetch origin rjhunter-remove-gec-from-highptmuon
git checkout -b rjhunter-remove-gec-from-highptmuon FETCH_HEAD
mkdir build
cd build
cmake -DSTANDALONE=ON -DSEQUENCES=hlt1_pp_default ..
make-
Run
/Allen --sequence hlt1_pp_default -
Copy that
Sequence.jsoninto stack replacing its pregenerated sequence e.g.cp Sequence.json /path/to/stack/Allen/configuration/pregenerated/hlt1_pp_default.json
make Allen/purge
make MooreAnalysis - Ran the job using the attached script.
Log files:
moore_options_file_result__run_did_not_end_+seg_fault.txt
moore_options_file_result__run_ended_+no_seg_faults.txt
moore_options_file_result__run_ended_+seg_fault.txt
Moreover,
Saw similar problems using upgrade_DC19_01_Z2mumuMU from TestFileDB. Options file attached below:
test_file_db_moore_options_file.py
test_file_db_moore_options_file_result__1_.txt