Horton

Thread

Commits

  1. doc: Fix section of functions age(xid) and mxid_age(xid)

  1. fix deprecation mention for age() and mxid_age()

    Bertrand Drouvot <bertranddrouvot.pg@gmail.com> — 2024-11-18T08:11:57Z

    Hi hackers,

48b5aa3143 mentioned age() and mxid_age() as deprecated. As per the discussion in
[1], it looks like that was not intentional.

Please find attached a patch to fix this on master and its version for the 17
stable branch backpatch. I think that's worth a backpatch as that could raise
concerns to read that the functions you rely on for monitoring are deprecated
and could generate questions like "will they be removed in PG 18?", "if so, what
is the supported/recommended way to calculate age of transaction id’s?".

Also it looks like that for PG 16, 7471230cc6, did put the functions description
in the "Control Data Functions". That looks unintentional too, so moving it to 
"Transaction ID and Snapshot Information Functions" section instead.

Please find attached the 3 patches.

[1]: https://www.postgresql.org/message-id/20231114013224.4z6oxa6p6va33rxr%40awork3.anarazel.de

Looking forward to your feedback,

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com

  2. Re: fix deprecation mention for age() and mxid_age()

    Michael Paquier <michael@paquier.xyz> — 2024-11-19T01:09:57Z

    On Mon, Nov 18, 2024 at 08:11:57AM +0000, Bertrand Drouvot wrote:
> Also it looks like that for PG 16, 7471230cc6, did put the functions description
> in the "Control Data Functions". That looks unintentional too, so moving it to 
> "Transaction ID and Snapshot Information Functions" section instead.
> 
> Please find attached the 3 patches.

Even with the monitoring query argument left aside, I agree that there
are more upsides to not list them as deprecated for now, based on the
reason that xid is still a data type in some catalogs, like pg_class's
relfrozenxid and relminmxid.  Note as well that we've used them in
vacuumdb.

In short, your suggestion makes sense to me.  Any thoughts or comments
from others?
--
Michael

  3. Re: fix deprecation mention for age() and mxid_age()

    Bertrand Drouvot <bertranddrouvot.pg@gmail.com> — 2024-11-19T06:52:40Z

    Hi,

On Tue, Nov 19, 2024 at 10:09:57AM +0900, Michael Paquier wrote:
> On Mon, Nov 18, 2024 at 08:11:57AM +0000, Bertrand Drouvot wrote:
> > Also it looks like that for PG 16, 7471230cc6, did put the functions description
> > in the "Control Data Functions". That looks unintentional too, so moving it to 
> > "Transaction ID and Snapshot Information Functions" section instead.
> > 
> > Please find attached the 3 patches.
> 
> Even with the monitoring query argument left aside, I agree that there
> are more upsides to not list them as deprecated for now, based on the
> reason that xid is still a data type in some catalogs, like pg_class's
> relfrozenxid and relminmxid.

Yeah, and those are most probably the main inputs for age(). In addition,
datminmxid and relminmxid are also xid and most probably the main inputs for
mxid_age().

> In short, your suggestion makes sense to me.

Thanks for the feedback!

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com

  4. Re: fix deprecation mention for age() and mxid_age()

    Michael Paquier <michael@paquier.xyz> — 2024-11-20T05:22:39Z

    On Tue, Nov 19, 2024 at 06:52:40AM +0000, Bertrand Drouvot wrote:
> Thanks for the feedback!

Done.  The section mistake in REL_16_STABLE was..  Interesting.
--
Michael

  5. Re: fix deprecation mention for age() and mxid_age()

    John Naylor <johncnaylorls@gmail.com> — 2024-12-10T23:54:23Z

    On Wed, Nov 20, 2024 at 12:23 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Tue, Nov 19, 2024 at 06:52:40AM +0000, Bertrand Drouvot wrote:
> > Thanks for the feedback!
>
> Done.  The section mistake in REL_16_STABLE was..  Interesting.

Hi, there is a CF entry for this -- is it ready to mark committed?

-- 
John Naylor
Amazon Web Services

  6. Re: fix deprecation mention for age() and mxid_age()

    Michael Paquier <michael@paquier.xyz> — 2024-12-11T00:01:24Z

    On Wed, Dec 11, 2024 at 06:54:23AM +0700, John Naylor wrote:
> Hi, there is a CF entry for this -- is it ready to mark committed?

Oops.  I've missed that there was an entry in CF 51.  Updated that
now.  Thanks for the poke.
--
Michael

  7. Re: fix deprecation mention for age() and mxid_age()

    Bertrand Drouvot <bertranddrouvot.pg@gmail.com> — 2024-12-11T07:04:16Z

    Hi,

On Wed, Dec 11, 2024 at 09:01:24AM +0900, Michael Paquier wrote:
> On Wed, Dec 11, 2024 at 06:54:23AM +0700, John Naylor wrote:
> > Hi, there is a CF entry for this -- is it ready to mark committed?
> 
> Oops.  I've missed that there was an entry in CF 51.  Updated that
> now.  Thanks for the poke.

D'oh I missed it too! 

> Thanks for the poke.

+1

Regards,

-- 
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com