I am running a transform query in hive 0.9.0 and trying to figure out why my script segfaults while run as part of a hive job but not part when run by itself. The library that is freaking out is libcrypto and is being called by the Python m2crypto module. That being said, I can run the same python script, on the same source data, and not in a transform script (on the same box!) with no issues. Confusing. So first of all, any reason this may be happening? If no takers there, how do you think we could trouble shoot this? Run strace on the script as it's being run in transform seems complicated. Just looking for ideas.
Thanks! John