Re: [HACKERS] Postgres 6.4.2 connection problem solved

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Andy Farrell <andy_farrell(at)itd(dot)sterling(dot)com>
Cc: PostGreSQL Hackers <pgsql-hackers(at)postgreSQL(dot)org>
Subject: Re: [HACKERS] Postgres 6.4.2 connection problem solved
Date: 1999-05-20 21:10:21
Message-ID: 3536.927234621@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Andy Farrell <andy_farrell(at)itd(dot)sterling(dot)com> writes:
> The machine we tried to run postgres on had the 'localhost' entry in
> its hosts file spelled incorrectly (i.e., 'localhosts'). After
> updating the hosts file, postgres ran fine.

That makes sense, if you were using TCP connection protocol rather
than a Unix-domain socket...

> I would have thought we would have recieved an error other than
> 'getprotobyname failed'.

I'll say. How the heck did it manage to get through gethostbyname()
and connect(), which are the routines that *should* have failed, and
then spit up at getprotobyname() (which should be nothing more than a
simple scan of /etc/protocols, and should certainly not care what is
in /etc/hosts)?

There is more than meets the eye here. If you have time, would you
restore /etc/hosts to its broken condition and trace through connectDB
a little more carefully? I would like to know what *really* went
wrong.

regards, tom lane

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 1999-05-20 21:15:54 Re: [HACKERS] 64 bit version on Solaris 7...
Previous Message Matthew C. Aycock 1999-05-20 19:54:10 64 bit version on Solaris 7...