Re: Transactions involving multiple postgres foreign servers, take 2 - Mailing list pgsql-hackers

From Masahiko Sawada
Subject Re: Transactions involving multiple postgres foreign servers, take 2
Date
Msg-id CAD21AoDa1tO1erBt1HGe=7ytXwyyKJbBiB9rtffr-WKH+SnV_w@mail.gmail.com
Whole thread Raw
In response to Re: Transactions involving multiple postgres foreign servers, take 2  (Fujii Masao <masao.fujii@oss.nttdata.com>)
Responses Re: Transactions involving multiple postgres foreign servers, take 2
List pgsql-hackers
On Fri, Jul 9, 2021 at 3:26 PM Fujii Masao <masao.fujii@oss.nttdata.com> wrote:
>
>
>
> On 2021/06/30 10:05, Masahiko Sawada wrote:
> > I've attached the new version patch that incorporates the comments
> > from Fujii-san and Ikeda-san I got so far.
>
> Thanks for updating the patches!
>
> I'm now reading 0001 and 0002 patches and wondering if we can commit them
> at first because they just provide independent basic mechanism for
> foreign transaction management.
>
> One question regarding them is; Why did we add new API only for "top" foreign
> transaction? Even with those patches, old API (CallSubXactCallbacks) is still
> being used for foreign subtransaction and xact_depth is still being managed
> in postgres_fdw layer (not PostgreSQL core). Is this intentional?

Yes, it's not needed for 2PC support and I was also concerned to add
complexity to the core by adding new API for subscriptions that are
not necessarily necessary for 2PC.

> As far as I read the code, keep using old API for foreign subtransaction doesn't
> cause any actual bug. But it's just strange and half-baked to manage top and
> sub transaction in the differenet layer and to use old and new API for them.

That's a valid concern. I'm really not sure what we should do here but
I guess that even if we want to support subscriptions we have another
API dedicated for subtransaction commit and rollback.

Regards,

--
Masahiko Sawada
EDB:  https://www.enterprisedb.com/



pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: [PATCH] Automatic HASH and LIST partition creation
Next
From: Tom Lane
Date:
Subject: Re: Support kerberos authentication for postgres_fdw