Date: Sun, 04 Nov 2012 11:45:15 +0100 From: Jan Danielsson <jan.m.daniels...@gmail.com>
When I think back, I recall noticing that I got brutally high throughput during the transfer; which raises the question: Is there any known "heavy load" bug lurking which causes the postgresql benchmark and high-throughput transfers to hang the system? Not that I know of, but perhaps we ought to have a PR to track these issues. When I ran pgbench on a 12-core (24-thread) system recently, I saw vaguely similar lock-up symptoms.