Re: Proposal: Implement failover on libpq connect level.

Victor Wagner wrote:

> It would just take a bit more time for client and a bit more load for
> server - to make sure that this connection is read-write by
> issuing
>
> show transaction_read_only
>
> statement before considering connection useful.

If the purpose of the feature is to wait for a failover to complete,
shouldn't it check for pg_is_in_recovery() rather than
transaction_read_only ?

That's because a database or user can be made read-only-on-connect
on an otherwise read-write instance by issuing
ALTER DATABASE dbname SET default_transaction_read_only TO on;
The same for a user with ALTER USER.

In that case, transaction_read_only would be OFF after connecting,
both on the master and on a slave, independantly of any failover
in progress or finished or not having occurred at all.

Best regards,
--
Daniel Vérité
PostgreSQL-powered mailer: http://www.manitou-mail.org
Twitter: @DanielVerite