| From: | Thom Brown <thom(at)linux(dot)com> |
|---|---|
| To: | Jelte Fennema <postgres(at)jeltef(dot)nl> |
| Cc: | Euler Taveira <euler(at)eulerto(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Allow specifying a dbname in pg_basebackup connection string |
| Date: | 2023-07-05 18:08:49 |
| Message-ID: | CAA-aLv6Zq062KUDQxVv8Om=uB=FP2qeAnFk4aztUBNBobgn0NQ@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, 5 Jul 2023 at 16:50, Jelte Fennema <postgres(at)jeltef(dot)nl> wrote:
>
> On Wed, 5 Jul 2023 at 16:01, Euler Taveira <euler(at)eulerto(dot)com> wrote:
> > One of the PgBouncer's missions is to be a transparent proxy.
> >
> > Sometimes you cannot reach out the database directly due to a security policy.
>
> Indeed the transparent proxy use case is where replication through
> pgbouncer makes sense. There's quite some reasons to set up PgBouncer
> like such a proxy apart from security policies. Some others that come
> to mind are:
> - load balancer layer of pgbouncers
> - transparent failovers
> - transparent database moves
>
> And in all of those cases its nice for a user to use a single
> connection string/hostname. Instead of having to think: Oh yeah, for
> backups, I need to use this other one.
Okay, understood. In that case, please remember to write changes to
the pg_basebackup docs page explaining how the dbname value is ignored
under normal usage.
Thom
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Soumyadeep Chakraborty | 2023-07-05 18:57:58 | Re: brininsert optimization opportunity |
| Previous Message | Thom Brown | 2023-07-05 17:54:53 | Re: Disabling Heap-Only Tuples |