From: Tom Lane [mailto:t...@sss.pgh.pa.us]
> Michael Paquier <mich...@paquier.xyz> writes:
> > On Tue, Jan 15, 2019 at 02:00:57AM +0000, Tsunakawa, Takayuki wrote:
> >> The original desire should have been the ability to connect to a
> >> primary or a standby. So, I think we should go back to the original
> >> thinking (and not complicate the feature), and create a read only
> >> GUC_REPORT variable, say, server_role, that identifies whether the
> >> server is a primary or a standby.
>
> > From the point of view of making sure that a client is really
> > connected to a primary or a standby, this is the best idea around.
>
> There are a couple of issues here:
> 1. Are you sure there are no use-cases for testing transaction_read_only
> as such?
I don't find any practical use case, but I won't object to leaving the current
target_session_attrs as-is. Alide from that, I think a parameter like PgJDBC's
is necessary, e.g., target_server_type = {primary | standby | prefer_standby},
which acts based on just the server role (primary or standby).
> 2. What will the fallback implementation be, when connecting to a server
> too old to have the variable you want?
One of the following:
1. "Unsupported" error. I'll take this.
2. libpq issues "SELECT pg_is_in_recovery()".
3. Blindly accepts the first successful connection.
Regards
Takayuki Tsunakawa
