2026-08-13 release announcement draft
Hi,
Please see attached draft of the 2026-08-13 release announcement, which
includes fixes both for the quarterly update release and the PG19 Beta.
Please check for accuracy, glaring omissions, confusing phrases, and typos.
I'll note that this was particularly interesting to pull together given
the number of fixes in this release, which was the most I've ever seen
(and by far). I tried to keep the list in its usual 20-25 range (and
keep in mind, the full list is available in the release notes) along
with my usual attempts to explain the fixes to a general audience, but
this was definitely a more challenging announcement to write up.
This is also to say a thanks to everyone for working to pull together
this update release!
Jonathan
Attachments:
20260813updaterelease.mdtext/markdown; charset=UTF-8; name=20260813updaterelease.mdDownload
"Jonathan S. Katz" <jkatz@postgresql.org> writes:
Please see attached draft of the 2026-08-13 release announcement, which
includes fixes both for the quarterly update release and the PG19 Beta.
Please check for accuracy, glaring omissions, confusing phrases, and typos.
One of the bigger-ticket bugs is:
* Fix a self-deadlock that could occur while replaying WAL generated by an older
minor version. This regression, introduced in the previous set of minor
releases, could cause a standby server following a primary running an older
minor version to become stuck.
Contrary to what you say at the top of the list, this doesn't affect
v18, only v14-v16. I think this one is definitely important enough
to mention, but maybe put it separately from the main list?
regards, tom lane
On 8/9/26 2:28 PM, Tom Lane wrote:
"Jonathan S. Katz" <jkatz@postgresql.org> writes:
Please see attached draft of the 2026-08-13 release announcement, which
includes fixes both for the quarterly update release and the PG19 Beta.
Please check for accuracy, glaring omissions, confusing phrases, and typos.One of the bigger-ticket bugs is:
* Fix a self-deadlock that could occur while replaying WAL generated by an older
minor version. This regression, introduced in the previous set of minor
releases, could cause a standby server following a primary running an older
minor version to become stuck.Contrary to what you say at the top of the list, this doesn't affect
v18, only v14-v16. I think this one is definitely important enough
to mention, but maybe put it separately from the main list?
The list was generated from my build of the v18 release notes, and this
was in there - though I found the commit message that did confirm it's
only v14-16.
Agree it deserves a special callout (I also remember this issue when it
was reported) - but do you want to remove it also from the 18.5 release
notes? (I checked release-18.sgml and couldn't find it).
Jonathan
"Jonathan S. Katz" <jkatz@postgresql.org> writes:
On 8/9/26 2:28 PM, Tom Lane wrote:
Contrary to what you say at the top of the list, this doesn't affect
v18, only v14-v16. I think this one is definitely important enough
to mention, but maybe put it separately from the main list?
The list was generated from my build of the v18 release notes, and this
was in there - though I found the commit message that did confirm it's
only v14-16.
That's an artifact of my release note process: the initial draft in
the latest stable branch documents everything, then after I copy that
text into the older branches I remove anything that's not relevant to
the latest branch.
Agree it deserves a special callout (I also remember this issue when it
was reported) - but do you want to remove it also from the 18.5 release
notes? (I checked release-18.sgml and couldn't find it).
I did remove it in a4c41bbcc, which is why you don't see it anymore.
regards, tom lane
On 8/9/26 3:03 PM, Tom Lane wrote:
"Jonathan S. Katz" <jkatz@postgresql.org> writes:
On 8/9/26 2:28 PM, Tom Lane wrote:
Contrary to what you say at the top of the list, this doesn't affect
v18, only v14-v16. I think this one is definitely important enough
to mention, but maybe put it separately from the main list?The list was generated from my build of the v18 release notes, and this
was in there - though I found the commit message that did confirm it's
only v14-16.That's an artifact of my release note process: the initial draft in
the latest stable branch documents everything, then after I copy that
text into the older branches I remove anything that's not relevant to
the latest branch.
I did remove it in a4c41bbcc, which is why you don't see it anymore.
Ah OK, good to know for the future (though there's still a nonzero
chance I trip over it should it happen again :)
What do you think of this framing:
===
This update fixes over 110 bugs that were reported in the last several
months.
This specific issue only affects PostgreSQL 14, 15, and 16, but we're
highlighting it in the release announcement due to its severity:
* Fix a self-deadlock that could occur while replaying WAL generated by
an older minor version. This regression, introduced in the previous set
of minor releases, could cause a standby server following a primary
running an older minor version to become stuck.
The remainder of issues listed below affect PostgreSQL 18. Some of these
issues may also affect other supported versions of PostgreSQL.
...
==
Thanks,
Jonathan
"Jonathan S. Katz" <jkatz@postgresql.org> writes:
What do you think of this framing:
===
This update fixes over 110 bugs that were reported in the last several
months.
This specific issue only affects PostgreSQL 14, 15, and 16, but we're
highlighting it in the release announcement due to its severity:
* Fix a self-deadlock that could occur while replaying WAL generated by
an older minor version. This regression, introduced in the previous set
of minor releases, could cause a standby server following a primary
running an older minor version to become stuck.
The remainder of issues listed below affect PostgreSQL 18. Some of these
issues may also affect other supported versions of PostgreSQL.
...
==
WFM, though personally I'd be more assertive in that last sentence:
"Many of these issues also affect older versions of PostgreSQL."
regards, tom lane
Hi.
https://www.postgresql.org/about/news/postgresql-186-1711-1615-1519-1424-and-19-beta-3-released-3365/
https://www.postgresql.org/docs/release/18.6/
SELECT DISTINCT t.oid::regclass, t.reltuples
FROM pg_class t JOIN pg_index i ON t.oid = i.indrelid
JOIN pg_class ic ON i.indexrelid = ic.oid
WHERE t.relhasindex AND ic.relam = 2742
This query appearance looks better in release notes than the announcement.
On 8/14/26 7:56 AM, jian he wrote:
Hi.
https://www.postgresql.org/about/news/postgresql-186-1711-1615-1519-1424-and-19-beta-3-released-3365/
https://www.postgresql.org/docs/release/18.6/SELECT DISTINCT t.oid::regclass, t.reltuples
FROM pg_class t JOIN pg_index i ON t.oid = i.indrelid
JOIN pg_class ic ON i.indexrelid = ic.oid
WHERE t.relhasindex AND ic.relam = 2742This query appearance looks better in release notes than the announcement.
The docstyles are different from some of the styles on the website. The
code on the website is also re-used for the emails, and is fairly
simplistic.
I think for this case, the most important part is people (or agents) can
copy and paste it.
Jonathan