Re: [GENERAL] logical replication in PG10 BETA - Mailing list pgsql-general

From Adrian Klaver
Subject Re: [GENERAL] logical replication in PG10 BETA
Date
Msg-id a2eb6495-8edd-9858-fcd3-8eda8de17598@aklaver.com
Whole thread Raw
In response to Re: [GENERAL] logical replication in PG10 BETA  (Igor Neyman <ineyman@perceptron.com>)
Responses Re: [GENERAL] logical replication in PG10 BETA  (Igor Neyman <ineyman@perceptron.com>)
List pgsql-general
On 05/25/2017 11:09 AM, Igor Neyman wrote:

> __________________________________________________________________________________________________________
>
> Adrian, thanks for trying to help.
>
> Even though the role I'm using (user=repl_user) has REPLICATION attribute,  I thought your question about .pgpass
filewas going to put me on a "right track" because I was not using/didn't have password file. 

I may have steered you wrong on this. The more think about it the more I
realize that the server will probably not read the .pgpass file.

> So, I created one proper password file (it works fine when I'm trying to connect through psql with no password).
Unfortunately,it didn't make any difference for CREATE SUBSCRIPTION. 

Two options come to mind to test whether the password is the problem:

1) Include the password in the connection string in CREATE SUBSCRIPTION.

2) Change the auth method in pg_hba.conf on the publisher server from
md5 to trust and reload the server.


Neither is optimal, still it is a starting point.

>
> I reported my problem as a bug (bug# 14669), but so far it doesn't seem to attract any interest.
>
> Regards,
> Igor Neyman
>


--
Adrian Klaver
adrian.klaver@aklaver.com


pgsql-general by date:

Previous
From: Sam Saffron
Date:
Subject: Re: [GENERAL] Why is posgres picking a suboptimal plan for this query?
Next
From: Igor Neyman
Date:
Subject: Re: [GENERAL] logical replication in PG10 BETA