adding setsid as a temporal solution for grid job
As discussed in ATR-20513, there is an issue on grid job for offline trigger reprocessing. On some sites, child processes of athenaHLT have the different PGID from the mother processes. They cannot be found from the mother process, and the job is finished immediately (0 events processed). We understood the issue is related to the option of the Singularity container. ADC experts will follow it up.
This MR is a temporal solution while waiting for the global solution. We can fix the PGID by hand by adding setsid command. A test job with this fix works fine: https://bigpanda.cern.ch/task/19870956/
Edited by Takuya Nobe