no pg_hba.conf entry for host

问题:

I get following error when I try to connect using DBI

DBI connect('database=chaosLRdb;host=192.168.0.1;port=5433','postgres',...) 
failed: FATAL:  no pg_hba.conf entry for host "192.168.0.1", user "postgres", database "chaosLRdb", SSL off

Here is my pg_hba.conf file:

# "local" is for Unix domain socket connections only
local   all         all                               md5
# IPv4 local connections:
host    all         all         127.0.0.1/32          md5
# IPv6 local connections:
host    all         all         ::1/128               md5

host    all         postgres    127.0.0.1/32          trust

host    all        postgres     192.168.0.1/32        trust

host    all        all         192.168.0.1/32        trust

host    all        all         192.168.0.1/128        trust

host    all        all         192.168.0.1/32        md5

host    chaosLRdb    postgres         192.168.0.1/32      md5
local    all        all         192.168.0.1/32        trust

My perl code is

#!/usr/bin/perl-w
use DBI;
use FileHandle;

print "Start connecting to the DB...\n";

@ary = DBI->available_drivers(true);
%drivers = DBI->installed_drivers();
my $dbh = DBI->connect("DBI:PgPP:database=chaosLRdb;host=192.168.0.1;port=5433", "postgres", "chaos123");

May I know what i miss here?

回答1:

In your pg_hba.conf file, I see some incorrect and confusing lines:

# fine, this allows all dbs, all users, to be trusted from 192.168.0.1/32
# not recommend because of the lax permissions
host    all        all         192.168.0.1/32        trust

# wrong, /128 is an invalid netmask for ipv4, this line should be removed
host    all        all         192.168.0.1/128       trust

# this conflicts with the first line
# it says that that the password should be md5 and not plaintext
# I think the first line should be removed
host    all        all         192.168.0.1/32        md5

# this is fine except is it unnecessary because of the previous line
# which allows any user and any database to connect with md5 password
host    chaosLRdb  postgres    192.168.0.1/32        md5

# wrong, on local lines, an IP cannot be specified
# remove the 4th column
local   all        all         192.168.0.1/32        trust

I suspect that if you md5'd the password, this might work if you trim the lines. To get the md5 you can use perl or the following shell script:

 echo -n 'chaos123' | md5sum
 > d6766c33ba6cf0bb249b37151b068f10  -

So then your connect line would like something like:

my $dbh = DBI->connect("DBI:PgPP:database=chaosLRdb;host=192.168.0.1;port=5433",
    "chaosuser", "d6766c33ba6cf0bb249b37151b068f10");

For more information, here's the documentation of postgres 8.X's pg_hba.conf file.

回答2:

If you can change this line:

host    all        all         192.168.0.1/32        md5

With this:

host    all        all         all                   md5

You can see if this solves the problem.

But another consideration is your postgresql port(5432) is very open to password attacks with hackers (maybe they can brute force the password). You can change your postgresql port 5432 to '33333' or another value, so they can't know this configuration.

回答3:

Your postgres server configuration seems correct

host    all         all         127.0.0.1/32          md5
host    all         all         192.168.0.1/32        trust

That should grant access from the client to the postgres server. So that leads me to believe the username / password is whats failing.

Test this by creating a specific user for that database

createuser -a -d -W -U postgres chaosuser

Then adjust your perl script to use the newly created user

my $dbh = DBI->connect("DBI:PgPP:database=chaosLRdb;host=192.168.0.1;port=5433", "chaosuser", "chaos123");

回答4:

To resolved this problem, you can try this.

first you have find out your pg_hba.conf and write :

local all all md5

after that restart pg server:

postgresql restart

or

sudo /etc/init.d/postgresql restart

回答5:

For those who have the similar problem trying to connect to local db and trying like
con = psycopg2.connect(database="my_db", user="my_name", password="admin"), try to pass the additional parameter, so the following saved me a day:
con = psycopg2.connect(database="my_db", user="my_name", password="admin", host="localhost")

回答6:

also check the PGHOST variable:

ECHO $PGHOST

to see if it matches the local machine name

回答7:

If you are getting an error like the one below:

OperationalError: FATAL:  no pg_hba.conf entry for host "your ipv6",
                  user "username", database "postgres", SSL off

then add an entry like the following, with your mac address.

host   all    all       [your ipv6]/128         md5

回答8:

BTW, in my case it was that I needed to specify the user/pwd in the url, not as independent properties, they were ignored and my OS user was used to connect

My config is in a WebSphere 8.5.5 server.xml file

<dataSource 
    jndiName="jdbc/tableauPostgreSQL" 
    type="javax.sql.ConnectionPoolDataSource">
    <jdbcDriver 
        javax.sql.ConnectionPoolDataSource="org.postgresql.ds.PGConnectionPoolDataSource" 
        javax.sql.DataSource="org.postgresql.ds.PGPoolingDataSource" 
        libraryRef="PostgreSqlJdbcLib"/>
    <properties 
        url="jdbc:postgresql://server:port/mydb?user=fred&amp;password=secret"/>
</dataSource>

This would not work and was getting the error:

<properties 
    user="fred"
    password="secret"
    url="jdbc:postgresql://server:port/mydb"/>

回答9:

For those who are getting this error in DBeaver the solution was found here at line:

@lcustodio on the SSL page, set SSL mode: require and either leave the SSL Factory blank or use the org.postgresql.ssl.NonValidatingFactory

Under Network -> SSL tab I checked the Use SLL checkbox and set Advance -> SSL Mode = require and it now works.