Skip to content

Bump PG minors for Citus 12.1 CI (14.24 / 15.19 / 16.15) - #243

Open
ibrahim halatci (ihalatci) wants to merge 1 commit into
bump-pg16-16.13-citus-12.1from
bump-pg-minors-citus-12.1-20260815
Open

Bump PG minors for Citus 12.1 CI (14.24 / 15.19 / 16.15)#243
ibrahim halatci (ihalatci) wants to merge 1 commit into
bump-pg16-16.13-citus-12.1from
bump-pg-minors-citus-12.1-20260815

Conversation

@ihalatci

@ihalatci ibrahim halatci (ihalatci) commented Aug 15, 2026

Copy link
Copy Markdown
Contributor

Bumps the PostgreSQL minors used by the Citus 12.1 CI image line to the
latest 2026-08-14 releases:

Major Was Now
14 14.19 14.24
15 15.14 15.19
16 16.13 16.15

All three are available on apt.postgresql.org for bookworm
(14.24-1.pgdg12+2, 15.19-1.pgdg12+2, 16.15-1.pgdg12+2).

Why this needs a CI look

This deliberately un-pins PG14 and PG15. Commit 98d1a12 held them back at
14.19 / 15.14 because the May-2026 coordinated minor set
(14.23 / 15.18 / 16.14) crashed PostgreSQL's own vanilla select_parallel
test on all three majors, and only PG16 was moved forward (to 16.13).

The point of this PR is to re-test that against the August set. If
select_parallel still crashes on PG14/PG15, the fallback is to keep this
line on the PG16-only bump (16.13 -> 16.15) and leave PG14/PG15 pinned.

The build-test-images run for this branch is green and published the dev
image set as image_suffix: -dev-b63fad9
(e.g. ghcr.io/citusdata/exttester:16.15-dev-b63fad9). Wire that suffix into
the Citus 12.1 test run to get the vanilla select_parallel signal — a green
image build here only proves the minors install, not that the tests pass.

Base branch

The 12.1 line has no long-lived base branch in this repo, so this targets
bump-pg16-16.13-citus-12.1, which is the current tip of that line.

Only circleci/images/PG_VERSIONS is touched.

Co-authored-by: Copilot 223556219+Copilot@users.noreply.github.com

… / 16.15)

PG14 14.19->14.24, PG15 15.14->15.19, PG16 16.13->16.15.

This intentionally un-pins PG14/PG15, which 98d1a12 held back at
14.19/15.14 because the May-2026 coordinated set (14.23/15.18/16.14)
crashed PostgreSQL's own vanilla select_parallel test. Re-testing
against the August set to see whether that regression is gone.

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant