Re: [PATCH] postgres_fdw connection caching - cause remote sessions linger till the local session exit

On Mon, Feb 1, 2021 at 12:29 PM Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com> wrote:
> On 2021/01/30 9:28, Bharath Rupireddy wrote:
> > On Sat, Jan 30, 2021 at 12:14 AM Fujii Masao
> > <masao(dot)fujii(at)oss(dot)nttdata(dot)com> wrote:
> >> + /*
> >> + * It doesn't make sense to show this entry in the output with a
> >> + * NULL server_name as it will be closed at the xact end.
> >> + */
> >> + continue;
> >>
> >> -1 with this change because I still think that it's more useful to list
> >> all the open connections.
> >
> > If postgres_fdw_get_connections doesn't have a "valid" column, then I
> > thought it's better not showing server_name NULL in the output.
>
> Or if we don't have strong reason to remove "valid" column,
> the current design is enough?

My only worry was that the statement from [1] "A cache flush should
not cause user-visible state changes." But the newly added function
postgres_fdw_get_connections is VOLATILE which means that the results
returned by postgres_fdw_get_connections() is also VOLATILE. Isn't
this enough, so that users will not get surprised with different
results in case invalidations occur within the server by the time they
run the query subsequent times and see different results than what
they saw in the first run?

Thoughts?

[1] https://www.postgresql.org/message-id/flat/2724627.1611886184%40sss.pgh.pa.us

With Regards,
Bharath Rupireddy.
EnterpriseDB: http://www.enterprisedb.com