segfault. which module to blame?

[Dan Parisien]

I am using shelve and a lot of threads to write to a dbm db. I implemented 
my own row-level locking with dictionaries and counters and when I run a 
test script that simulates extreme conditions (1500 threads, 10 writes 
each) it segfaults 1/100 times it runs.

That kinda sucks, don't you think? I'm wondering if someone who has 
experience could point me to which module is at fault: threading? or 
shelve+dbm? or neither? :) Or maybe some technique on how to crawl through 
the core file (I'm on mandrake gnu/linux 7.2 python 2.0.1)

I'm not expecting a correct answer, just a point in the right directions

Thank you :)
Dan