Pat Maddox <pat.maddox@gmail.com> writes:
> On Thu, Mar 5, 2009 at 10:34 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Odd. �Try strace'ing the failing psql session, and send us the last few
>> dozen lines of the output.
> http://pastie.org/private/y7cbpuamemxawmhh8hdna is the strace output
So the relevant info is
sendto(3, "\0\0\0*\0\3\0\0user\0enc\0database\0yachti"..., 42, 0, NULL, 0) = 42
rt_sigaction(SIGPIPE, {SIG_DFL}, {SIG_IGN}, 8) = 0
poll([{fd=3, events=POLLIN|POLLERR, revents=POLLIN}], 1, -1) = 1
recvfrom(3, "E\0\0\0\225SFATAL\0C28000\0Mno pg_hba.co"..., 16384, 0, NULL, NULL) = 150
write(2, "psql: ", 6psql: ) = 6
close(3) = 0
exit_group(2) = ?
Which looks to me like the server is correctly returning a rejection
(I assume there's a "no pg_hba.conf entry..." error showing up in the
server log too?) and then libpq is somehow falling down at converting
that to a displayable message --- psql seems to be getting told
correctly that the connection failed, but it's getting an empty string
instead of the desired error message.
In short, you've got a broken libpq (or possibly a broken psql
executable) on that machine. I'd try reinstalling the postgres stuff
before spending additional time on it. If it still fails then it would
be time to inquire into configuration options and so forth used to build
the postgres code for that machine.
regards, tom lane