Hugo Osvaldo Barrera <hugo@barrera.io> writes:
> The included patch fixes the issue and postgres has been working fine for me
> with it. See this thread[2] for details.
Hm, I was wondering whether copying sizeof(struct sockaddr_storage) might
not be a risky choice. However, the conclusion on the -hackers thread
http://www.postgresql.org/message-id/flat/23855.1424136475@sss.pgh.pa.us
seems to be that we're going to eliminate the entire issue by just ripping
out this code section, lock stock n barrel.
Thanks for confirming that this fixes it for you, though --- at least
we now have a certain diagnosis of the problem.
regards, tom lane