On Tue, Jun 6, 2023 at 4:11 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Tue, Jun 06, 2023 at 10:00:00AM +0530, Bharath Rupireddy wrote:
> > +1 for deduplicating find_in_log. How about deduplicating advance_wal
> > too so that 019_replslot_limit.pl, 033_replay_tsp_drops.pl,
> > 035_standby_logical_decoding.pl and 001_stream_rep.pl can benefit
> > immediately?
>
> As in a small wrapper for pg_switch_wal() that generates N segments at
> will?
Yes. A simpler way of doing it would be to move advance_wal() in
019_replslot_limit.pl to Cluster.pm, something like the attached. CI
members don't complain with it
https://github.com/BRupireddy/postgres/tree/add_a_function_in_Cluster.pm_to_generate_WAL.
Perhaps, we can name it better instead of advance_wal, say
generate_wal or some other?
> I don't see why we could not do that if it proves useful in the
> long run.
Besides the beneficiaries listed above, the test case added by
https://commitfest.postgresql.org/43/3663/ can use it. And, the
test_table bits in 020_pg_receivewal.pl can use it (?).
--
Bharath Rupireddy
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com