RE: Add client connection check during the execution of the query

From: Thomas Munro <thomas(dot)munro(at)gmail(dot)com>
> > Following PostmasterIsAlive(), maybe it's better to name it as
> pq_connection_is_alive() pq_client_is_alive(), or pq_frontend_is_alive() like
> the pqcomm.c's head comment uses the word frontend?
>
> I think it's OK, because it matches the name of the GUC. I'm more concerned
> about the name of the GUC. Will we still be happy with this name if a future
> releases sends a heartbeat message? I think that is still OK, so I'm happy with
> these names for now, but if someone has a better name, please speak up very
> soon.

OK, agreed.

> > (4)
> > I think the new GUC works for walsender as well. If so, how do we explain
> the relationship between the new GUC and wal_receiver_timeout and
> recommend the settings of them?
>
> No, it only works while executing a query. (Is there something in logical
> decoding, perhaps, that I have failed to consider?)

When I saw the following code, I thought the new GUC takes effect at the same time as wal\sender_timeout. But they don't become effective simultaneously. The new GUC is effective before the logical replication starts. And wal_sender_timeout takes effect after the physical/logical replication starts. So, there's no problem.

[PostgresMain]
if (am_walsender)
{
if (!exec_replication_command(query_string))
exec_simple_query(query_string);
}
else
exec_simple_query(query_string);

The patch looks committable to me.

> PS The "from" headers in emails received from Fujitsu seems to have the
> names stripped, somewhere in the tubes of the internet. I see the full version
> when people from Fujitsu quote other people from Fujitsu.
> I copied one of those into the commit message, complete with its magnificent
> kanji characters (perhaps these are the cause of the filtering?), and I hope
> that's OK with you.

Certainly, it seems that only my email address appears in the sender field on the mailer. I'll check my Outlook settings.

Regards
Takayuki Tsunakawa