[pgpool-general: 7440] Re: Supported python async drivers

[Anton Melser]

Thanks very much for investigating that!. I experienced something very
similar to https://github.com/MagicStack/asyncpg/issues/573, which was also
experienced by a few others. I will try and set up a basic reproduction use
case. I will note that it does appear that *all* of us are using the
Bitnami postgresql-ha images/helm chart... So it may actually be a special
config they use, or possibly related to having it in a docker image... I'll
investigate further and report back.

Thanks again for your quick reply on this!

On Wed, 17 Mar 2021 at 21:20, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:

> >> Hi,
> >>
> >> I am using pgpool2 for failover (and maybe later load balancing, it's a
> >> long story...) and am starting to migrate some of my code to use async.
> >> Many async projects that support postgres I have seen do so via asyncpg.
> >> asyncpg doesn't seem to work at all with pgpool2 (at least with my
> current
> >> pgpool config, which works perfectly synchronously with psycopg2). Basic
> >> initial testing suggests aiopg does support pgpool2, at least for what I
> >> tested it with. This is probably because it is based on psycopg2-binary.
> >>
> >> Unfortunately almost all of the projects I have been looking at using
> that
> >> require async postgres only support asyncpg (Broadcaster,
> TortoiseORM,...),
> >> not aiopg.
> >>
> >> Is there an officially supported async driver for python? Is one
> planned?
> >
> > I am not familiar with "async" nor "asyncpg". To understand what they
> > are doing, it would be helpful to config Pgpool-II with
> > log_client_messages = on and share the pgpool log.
>
> I have tried with asyncpg myself. (see attached python
> program. Forgive me if there's something wrong. This is my first
> python script:-) With log_client_messages = on, and log_connections =
> on it seems Pgpool-II works fine.
>
> 2021-03-17 22:15:33: child pid 28311: LOG:  new connection received
> 2021-03-17 22:15:33: child pid 28311: DETAIL:  connecting host=127.0.0.1
> port=33452
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Parse message from
> frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: DETAIL:  statement:
> "__asyncpg_stmt_1__", query: "SELECT 1;"
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 0 backend pid:
> 28670 statement: SELECT version()
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 2 backend pid:
> 28669 statement: Parse: SELECT 1;
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Describe message from
> frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: DETAIL:  statement:
> "__asyncpg_stmt_1__"
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 2 backend pid:
> 28669 statement: D message
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Sync message from frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Bind message from frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: DETAIL:  portal: "", statement:
> "__asyncpg_stmt_1__"
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 2 backend pid:
> 28669 statement: Bind: SELECT 1;
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Execute message from
> frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: DETAIL:  portal: ""
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 2 backend pid:
> 28669 statement: Execute: SELECT 1;
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Sync message from frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  Terminate message from
> frontend.
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 0 backend pid:
> 28670 statement:  DISCARD ALL
> 2021-03-17 22:15:33: [unknown] pid 28311: LOG:  DB node id: 2 backend pid:
> 28669 statement:  DISCARD ALL
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese:http://www.sraoss.co.jp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20210317/2bc63f9b/attachment.htm>