Обсуждение: 2023-08-10 release announcement draft
Hi, Attached is the release announcement draft for the 2023-08-10 update release, which also includes the release of PostgreSQL 16 Beta 3. Please provide your feedback no later than August 10, 2023 0:00 AoE[1]. Thanks, Jonathan [1] https://en.wikipedia.org/wiki/Anywhere_on_Earth
Вложения
On Tue, 8 Aug 2023 at 13:15, Jonathan S. Katz <jkatz@postgresql.org> wrote: > Attached is the release announcement draft for the 2023-08-10 update > release, which also includes the release of PostgreSQL 16 Beta 3. Thanks for drafting this. > * Fix a performance regression when running concurrent > [`COPY`](https://www.postgresql.org/docs/16/sql-copy.html) statements on a > single table. I think this is still outstanding. A bit of work has been done for the int parsing regression but it seems there's still a performance regression when running multiple COPYs on the same table, per [1]. > or an previous major version of PostgreSQL, you will need to use a strategy "a previous". David [1] https://postgr.es/m/CAD21AoC0GvCEZbDreoAcj=i0LjNCQePQbh_cxCuBKezYgYwmTA@mail.gmail.com
On 8/7/23 9:45 PM, David Rowley wrote: >> * Fix a performance regression when running concurrent >> [`COPY`](https://www.postgresql.org/docs/16/sql-copy.html) statements on a >> single table. > > I think this is still outstanding. A bit of work has been done for the > int parsing regression but it seems there's still a performance > regression when running multiple COPYs on the same table, per [1]. Hm, the open item was closed[1] -- was that premature, or is this a new issue (have not yet read the thread you referenced)? >> or an previous major version of PostgreSQL, you will need to use a strategy > > "a previous". Thanks for the catch -- fixed locally. Jonathan [1] https://wiki.postgresql.org/index.php?title=PostgreSQL_16_Open_Items#resolved_before_16beta3
Вложения
On Tue, 8 Aug 2023 at 13:49, Jonathan S. Katz <jkatz@postgresql.org> wrote: > > On 8/7/23 9:45 PM, David Rowley wrote: > > >> * Fix a performance regression when running concurrent > >> [`COPY`](https://www.postgresql.org/docs/16/sql-copy.html) statements on a > >> single table. > > > > I think this is still outstanding. A bit of work has been done for the > > int parsing regression but it seems there's still a performance > > regression when running multiple COPYs on the same table, per [1]. > > Hm, the open item was closed[1] -- was that premature, or is this a new > issue (have not yet read the thread you referenced)? I closed it thinking that enough had been done to resolve the performance regression. In the linked thread, Sawadasan shows that that's not the case. So, yes, premature. I've reverted the change to the open items list now. David
On 8/7/23 9:53 PM, David Rowley wrote: > On Tue, 8 Aug 2023 at 13:49, Jonathan S. Katz <jkatz@postgresql.org> wrote: >> >> On 8/7/23 9:45 PM, David Rowley wrote: >> >>>> * Fix a performance regression when running concurrent >>>> [`COPY`](https://www.postgresql.org/docs/16/sql-copy.html) statements on a >>>> single table. >>> >>> I think this is still outstanding. A bit of work has been done for the >>> int parsing regression but it seems there's still a performance >>> regression when running multiple COPYs on the same table, per [1]. >> >> Hm, the open item was closed[1] -- was that premature, or is this a new >> issue (have not yet read the thread you referenced)? > > I closed it thinking that enough had been done to resolve the > performance regression. In the linked thread, Sawadasan shows that > that's not the case. So, yes, premature. I've reverted the change to > the open items list now. Got it. I reverted it as well from the release announcement. Reattaching with the clean copy. (Aside: I'm super excited for this PG16 improvement + fixed regression, as lately I've had to do some bulk imports on a single table that could really benefit from this :) Jonathan
Вложения
Op 8/8/23 om 03:15 schreef Jonathan S. Katz: > > Please provide your feedback no later than August 10, 2023 0:00 AoE[1]. 'You us' should be 'You use' (2x) Erik
On 8/8/23 1:30 AM, Erik Rijkers wrote: > Op 8/8/23 om 03:15 schreef Jonathan S. Katz: >> >> Please provide your feedback no later than August 10, 2023 0:00 AoE[1]. > > 'You us' should be > 'You use' > (2x) It should actually be just "Use" -- but I've fixed both instances. Thanks! Jonathan
Вложения
On Mon, Aug 7, 2023 at 9:15 PM Jonathan S. Katz <jkatz@postgresql.org> wrote: > > Hi, > > Attached is the release announcement draft for the 2023-08-10 update > release, which also includes the release of PostgreSQL 16 Beta 3. > > Please provide your feedback no later than August 10, 2023 0:00 AoE[1]. > > Thanks, > > Jonathan > > [1] https://en.wikipedia.org/wiki/Anywhere_on_Earth "Users who have skipped one or more update releases may need to run additional, post-update steps; " The comma should be removed. "please see the release notes for earlier versions for details." Use of 'for' twice is grammatically incorrect; I am partial to "please see the release notes from earlier versions for details." but could also see "please see the release notes for earlier versions to get details." Robert Treat https://xzilla.net
On Mon, Aug 07, 2023 at 10:03:44PM -0400, Jonathan S. Katz wrote: > Fixes in PostgreSQL 16 Beta 3 > ----------------------------- > > The following includes fixes included in PostgreSQL 16 Beta 3: With both "includes" and "included" there, this line reads awkwardly to me. I'd just delete the line, since the heading has the same information. > * Add the `\drg` command to `psql` to display information about role grants. > * Add timeline ID to filenames generated with `pg_waldump --save-fullpage`. > * Fix crash after a deadlock occurs in a parallel `VACUUM` worker. > > Please see the [release notes](https://www.postgresql.org/docs/16/release-16.html) > for a complete list of new and changed features: > > [https://www.postgresql.org/docs/16/release-16.html](https://www.postgresql.org/docs/16/release-16.html)
On 8/9/23 1:04 AM, Noah Misch wrote: > On Mon, Aug 07, 2023 at 10:03:44PM -0400, Jonathan S. Katz wrote: >> Fixes in PostgreSQL 16 Beta 3 >> ----------------------------- >> >> The following includes fixes included in PostgreSQL 16 Beta 3: > > With both "includes" and "included" there, this line reads awkwardly to me. > I'd just delete the line, since the heading has the same information. I took your suggestion. Thanks! Jonathan
Вложения
On 8/8/23 11:13 PM, Robert Treat wrote: > "Users who have skipped one or more update releases may need to run > additional, post-update steps; " > > The comma should be removed. > > "please see the release notes for earlier versions for details." > > Use of 'for' twice is grammatically incorrect; I am partial to "please > see the release notes from earlier versions for details." but could > also see "please see the release notes for earlier versions to get > details." Interestingly, I think this language has been unmodified for awhile. Upon reading it, I agree, and took your suggestions. Thanks, Jonathan
Вложения
Hi,
Sorry, join late after the release note is out.
> * An out-of-memory error from JIT will now cause a PostgreSQL `FATAL` error instead of a C++ exception.
I assume this statement is related to this fix[1].
I think an OOM error from JIT causing a PostgreSQL `FATAL` error is the actual behavior before the fix.
What the fix does is to bring exceptions back to the c++ world when OOM happens outside of LLVM.
Thanks,
Donghang Lin
On Wed, Aug 9, 2023 at 7:10 PM Jonathan S. Katz <jkatz@postgresql.org> wrote:
On 8/8/23 11:13 PM, Robert Treat wrote:
> "Users who have skipped one or more update releases may need to run
> additional, post-update steps; "
>
> The comma should be removed.
>
> "please see the release notes for earlier versions for details."
>
> Use of 'for' twice is grammatically incorrect; I am partial to "please
> see the release notes from earlier versions for details." but could
> also see "please see the release notes for earlier versions to get
> details."
Interestingly, I think this language has been unmodified for awhile.
Upon reading it, I agree, and took your suggestions.
Thanks,
Jonathan