Python error on one of the machines

FCCSW
1

Hello,
I am merely writing this here to report a problem on machine 753.
I have had to run an analysis this morning on a large file and when using the analysis.py file from FCCAnalyses I got the following error :
Load cxx analyzers …
edm4hep <cppyy.gbl.edm4hep.ReconstructedParticleData object at 0x4e76150>
podio <cppyy.gbl.podio.ObjectID object at 0x60d8f90>
fccana <C++ overload “dummyLoader” at 0x7f6c20a6ff10>
done
*** Error in `python’: double free or corruption (fasttop): 0x00000000078f2520 ***
======= Backtrace: =========

followed by (lengthy) precisions on the problem that seemed to point towards a version error in python. Having researched the problem on internet, it seemed indeed to be a problem on the distant machine and not with my code. I logged onto a different machine and it worked.
Hope that the problem with the machine can be fixed,
Mariette

2

thanks @mjolly!
Could you confirm if you issue is reproducible on machine 753?
Best,
Clement

3

I believe so, I don’t know how to choose the machine I’m connecting on to check. You are very welcome to use my codes to check, I was running it using the analysis.py that is at /afs/cern.ch/user/m/mjolly/public/Data/ and the problem should occur regardless of the root file given its nature. If not, I’ll be happy to provide you with more info on the problem.