"Moshe Ben-Shoham" <mosheb(at)nexperience(dot)com> writes:
> I believe there are two problems here: one is the process 2604 exiting
> with code 128, and the other one is the re-initialization failure with
> the following fatal message: "BST FATAL: pre-existing shared memory
> block is still in use".
The second problem is most likely going to be resolved by this recent fix:
http://archives.postgresql.org/pgsql-committers/2009-05/msg00085.php
I have no idea what is provoking an exit(128) though.
regards, tom lane