Thread: 'no pg_hba.conf entry for host "[local]", user "postgres", database "postgres"'...

When I try to run

% psql

as the postgres user, I get the error

psql: FATAL:  no pg_hba.conf entry for host "[local]", user "postgres", database "postgres", SSL off

I don't understand this error, since the pg_hba.conf file includes the line:

local all all ident sameuser

What could explain this behavior?

Incidentally, is there a general way to get positive confirmation that a particular pg_hba.conf has been re-read upon bouncing the server with

kill -HUP <PID>

?

TIA!

kynn




On Mon, Apr 13, 2009 at 2:01 PM, Kynn Jones <kynnjo@gmail.com> wrote:
> When I try to run
> % psql
> as the postgres user, I get the error
> psql: FATAL:  no pg_hba.conf entry for host "[local]", user "postgres",
> database "postgres", SSL off
>
> I don't understand this error, since the pg_hba.conf file includes the line:
> local all all ident sameuser
> What could explain this behavior?

What do all the lines look like together?

On Mon, Apr 13, 2009 at 4:54 PM, Scott Marlowe <scott.marlowe@gmail.com> wrote:
On Mon, Apr 13, 2009 at 2:01 PM, Kynn Jones <kynnjo@gmail.com> wrote:
> When I try to run
> % psql
> as the postgres user, I get the error
> psql: FATAL:  no pg_hba.conf entry for host "[local]", user "postgres",
> database "postgres", SSL off
>
> I don't understand this error, since the pg_hba.conf file includes the line:
> local all all ident sameuser
> What could explain this behavior?

What do all the lines look like together?

The uncommented lines are:

local   all     yours_truly        password
host    all     yours_truly        0.0.0.1/0            md5
host    all     yours_truly        ::1/128              md5
local  all     all             ident sameuser

Thanks!

Kynn
 

Kynn Jones <kynnjo@gmail.com> writes:
> On Mon, Apr 13, 2009 at 4:54 PM, Scott Marlowe <scott.marlowe@gmail.com>wrote:
>> On Mon, Apr 13, 2009 at 2:01 PM, Kynn Jones <kynnjo@gmail.com> wrote:
>>> as the postgres user, I get the error
>>> psql: FATAL:  no pg_hba.conf entry for host "[local]", user "postgres",
>>> database "postgres", SSL off

>> What do all the lines look like together?

> The uncommented lines are:

> local   all     yours_truly        password
> host    all     yours_truly        0.0.0.1/0            md5
> host    all     yours_truly        ::1/128              md5
> local  all     all             ident sameuser

That connection should certainly match the last line, and maybe the first
too; but it didn't find a match at all.  I conclude that this file isn't
actually what's driving the postmaster --- you are looking at the wrong
file, or you changed it and forgot to SIGHUP the postmaster, or you're
connecting to some other postmaster altogether, or some other mistake
not visible in the terms of this discussion.

            regards, tom lane

Kynn Jones <kynnjo@gmail.com> writes:
> Is there a general way to get positive confirmation that a particular
> pg_hba.conf has been re-read upon bouncing the server with
> kill -HUP <PID>

Well, you could strace the postmaster and see what files it reads
in response to the signal.

            regards, tom lane



On Tue, Apr 14, 2009 at 12:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Kynn Jones <kynnjo@gmail.com> writes:
> On Mon, Apr 13, 2009 at 4:54 PM, Scott Marlowe <scott.marlowe@gmail.com>wrote:
>> On Mon, Apr 13, 2009 at 2:01 PM, Kynn Jones <kynnjo@gmail.com> wrote:
>>> as the postgres user, I get the error
>>> psql: FATAL:  no pg_hba.conf entry for host "[local]", user "postgres",
>>> database "postgres", SSL off

>> What do all the lines look like together?

> The uncommented lines are:

> local   all     yours_truly        password
> host    all     yours_truly        0.0.0.1/0            md5
> host    all     yours_truly        ::1/128              md5
> local  all     all             ident sameuser

That connection should certainly match the last line, and maybe the first
too; but it didn't find a match at all.  I conclude that this file isn't
actually what's driving the postmaster --- you are looking at the wrong
file, or you changed it and forgot to SIGHUP the postmaster, or you're
connecting to some other postmaster altogether, or some other mistake
not visible in the terms of this discussion.


Is there a general way to get positive confirmation that a particular pg_hba.conf has been re-read upon bouncing the server with

kill -HUP <PID>

?

TIA!

Kynn

 
Kynn Jones escribió:

> Is there a general way to get positive confirmation that a particular
> pg_hba.conf has been re-read upon bouncing the server with

SHOW hba_file;

--
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.



On Tue, Apr 14, 2009 at 2:53 PM, Alvaro Herrera <alvherre@commandprompt.com> wrote:
Kynn Jones escribió:

> Is there a general way to get positive confirmation that a particular
> pg_hba.conf has been re-read upon bouncing the server with

SHOW hba_file;

That's handy.  Thanks!

Kynn