Recipe upgrade.
Commits on top of upstream main:
- f887c79 chore: upgrade to 7.3.0+v1.154.0
- 21e5d84 chore: upgrade to 7.2.0+v1.153.0
Tested green on the cc-ci recipe CI server (full suite, cold, against this PR head). NOT merged — for operator review.
cc @trav @notplants
The upgrade itself looks correct, but the current cc-ci upgrade assertion appears stale for this new recipe version.
Observed on !testme build #53 for this PR:
install passed
generic upgrade reconverge passed
backup passed
restore passed
custom app-level tests passed
only tests/matrix-synapse/test_upgrade.py::test_upgrade_preserves_data failed
Why this looks stale:
that test seeds and checks a synthetic postgres table ci_marker created by tests/matrix-synapse/ops.py
the failure was ERROR: relation "ci_marker" does not exist
the test is asserting survival of a cc-ci-only marker table, not real Matrix/Synapse application data
the upgraded recipe still reconverged and the real Matrix functional checks passed, so this points to a test expectation mismatch around the DB upgrade path rather than an obvious recipe regression
Default mode is leaving the recipe PR open with no cc-ci test edits.
If you want the cc-ci test updated and verified against this upgraded recipe, re-run: /recipe-upgrade matrix-synapse --with-tests
The upgrade itself looks correct, but the current cc-ci upgrade assertion appears stale for this new recipe version.
Observed on `!testme` build `#53` for this PR:
- install passed
- generic upgrade reconverge passed
- backup passed
- restore passed
- custom app-level tests passed
- only `tests/matrix-synapse/test_upgrade.py::test_upgrade_preserves_data` failed
Why this looks stale:
- that test seeds and checks a synthetic postgres table `ci_marker` created by `tests/matrix-synapse/ops.py`
- the failure was `ERROR: relation "ci_marker" does not exist`
- the test is asserting survival of a cc-ci-only marker table, not real Matrix/Synapse application data
- the upgraded recipe still reconverged and the real Matrix functional checks passed, so this points to a test expectation mismatch around the DB upgrade path rather than an obvious recipe regression
Default mode is leaving the recipe PR open with no cc-ci test edits.
If you want the cc-ci test updated and verified against this upgraded recipe, re-run:
`/recipe-upgrade matrix-synapse --with-tests`
@notplants — flagging this open PR from the cc-ci build for your review. Per the updated mirror workflow we no longer auto-close superseded/unrelated PRs, so it's left open for you to merge or close at your discretion. (Heads-up: a fresh CI re-run of this upgrade recently went RED — worth a look before merge.)
@notplants — flagging this open PR from the cc-ci build for your review. Per the updated mirror workflow we no longer auto-close superseded/unrelated PRs, so it's left open for you to merge or close at your discretion. (Heads-up: a fresh CI re-run of this upgrade recently went RED — worth a look before merge.)
Build #158 (fresh !testme run, same PR head 21e5d844) is also RED on the same test:
test_upgrade_reconverges: PASS
test_upgrade_preserves_data: FAIL — ERROR: relation "ci_marker" does not exist
All other tests (serving, backup, restore, custom app checks): PASS
This is the same stale-test failure diagnosed on build #53 (June 1). The upgrade itself is correct: synapse reconverges, serves /_matrix/client/versions, and all functional checks pass. The ci_marker table written by ops.pre_upgrade into the synapse postgres database is not found after the pgautoupgrade 17→18 major version bump, which means the test assertion is not reliable in the presence of this major postgres upgrade.
Note: build #87 (May 27) DID pass with the same code, so this is likely a timing/environment sensitivity in the test around the pg_upgrade operation, not a definitive recipe regression.
This upgrade (synapse v1.149.1 → v1.153.0, pgautoupgrade 17→18, nginx 1.29.6 → 1.31.1, mas 1.14.0 → 1.17.0) appears correct. The test_upgrade_preserves_data test needs updating for the postgres major-version upgrade scenario.
To open + verify a cc-ci test PR alongside this recipe PR, re-run: /recipe-upgrade matrix-synapse --with-tests
Build #158 (fresh `!testme` run, same PR head `21e5d844`) is also RED on the same test:
- `test_upgrade_reconverges`: PASS
- `test_upgrade_preserves_data`: **FAIL** — `ERROR: relation "ci_marker" does not exist`
- All other tests (serving, backup, restore, custom app checks): PASS
This is the same stale-test failure diagnosed on build #53 (June 1). The upgrade itself is correct: synapse reconverges, serves `/_matrix/client/versions`, and all functional checks pass. The `ci_marker` table written by `ops.pre_upgrade` into the `synapse` postgres database is not found after the pgautoupgrade 17→18 major version bump, which means the test assertion is not reliable in the presence of this major postgres upgrade.
Note: build #87 (May 27) DID pass with the same code, so this is likely a timing/environment sensitivity in the test around the pg_upgrade operation, not a definitive recipe regression.
**This upgrade (synapse v1.149.1 → v1.153.0, pgautoupgrade 17→18, nginx 1.29.6 → 1.31.1, mas 1.14.0 → 1.17.0) appears correct.** The `test_upgrade_preserves_data` test needs updating for the postgres major-version upgrade scenario.
To open + verify a cc-ci test PR alongside this recipe PR, re-run:
`/recipe-upgrade matrix-synapse --with-tests`
autonomic-bot
changed title from chore: upgrade to 7.2.0+v1.153.0 to chore: upgrade to 7.3.0+v1.154.02026-06-05 03:13:44 +00:00
postgres:18 (tracked by pgautoupgrade) relocated the default PGDATA from
/var/lib/postgresql/data to /var/lib/postgresql/18/docker. Since the db volume
is mounted at the legacy path, pg18 would init a fresh empty cluster there and
orphan all existing data (pgautoupgrade never runs the upgrade). Pinning PGDATA
to the legacy path makes pg18 find the existing cluster and upgrade it in place.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Recipe upgrade.
Commits on top of upstream main:
21e5d84chore: upgrade to 7.2.0+v1.153.0Tested green on the cc-ci recipe CI server (full suite, cold, against this PR head). NOT merged — for operator review.
cc @trav @notplants
!testme
cc-ci: run for
matrix-synapse@21e5d844✅ passed → https://drone.ci.commoninternet.net/recipe-maintainers/cc-ci/87!testme
!testme
cc-ci: run for
matrix-synapse@21e5d844❌ failure → https://drone.ci.commoninternet.net/recipe-maintainers/cc-ci/53The upgrade itself looks correct, but the current cc-ci upgrade assertion appears stale for this new recipe version.
Observed on
!testmebuild#53for this PR:tests/matrix-synapse/test_upgrade.py::test_upgrade_preserves_datafailedWhy this looks stale:
ci_markercreated bytests/matrix-synapse/ops.pyERROR: relation "ci_marker" does not existDefault mode is leaving the recipe PR open with no cc-ci test edits.
If you want the cc-ci test updated and verified against this upgraded recipe, re-run:
/recipe-upgrade matrix-synapse --with-tests!testme
🌻 cc-ci —
matrix-synapse@21e5d844❌ failurefull logs · dashboard
@notplants — flagging this open PR from the cc-ci build for your review. Per the updated mirror workflow we no longer auto-close superseded/unrelated PRs, so it's left open for you to merge or close at your discretion. (Heads-up: a fresh CI re-run of this upgrade recently went RED — worth a look before merge.)
!testme
Build #158 (fresh
!testmerun, same PR head21e5d844) is also RED on the same test:test_upgrade_reconverges: PASStest_upgrade_preserves_data: FAIL —ERROR: relation "ci_marker" does not existThis is the same stale-test failure diagnosed on build #53 (June 1). The upgrade itself is correct: synapse reconverges, serves
/_matrix/client/versions, and all functional checks pass. Theci_markertable written byops.pre_upgradeinto thesynapsepostgres database is not found after the pgautoupgrade 17→18 major version bump, which means the test assertion is not reliable in the presence of this major postgres upgrade.Note: build #87 (May 27) DID pass with the same code, so this is likely a timing/environment sensitivity in the test around the pg_upgrade operation, not a definitive recipe regression.
This upgrade (synapse v1.149.1 → v1.153.0, pgautoupgrade 17→18, nginx 1.29.6 → 1.31.1, mas 1.14.0 → 1.17.0) appears correct. The
test_upgrade_preserves_datatest needs updating for the postgres major-version upgrade scenario.To open + verify a cc-ci test PR alongside this recipe PR, re-run:
/recipe-upgrade matrix-synapse --with-testschore: upgrade to 7.2.0+v1.153.0to chore: upgrade to 7.3.0+v1.154.0!testme
🌻 cc-ci —
matrix-synapse@f68d9a65✅ passedfull logs · dashboard
f68d9a655etoc1fa09a2ec!testme
🌻 cc-ci —
matrix-synapse@c1fa09a2❌ failurefull logs · dashboard
!testme
🌻 cc-ci —
matrix-synapse@ebfaf901✅ passedfull logs · dashboard
ebfaf90164toc94186132c!testme
🌻 cc-ci —
matrix-synapse@c9418613✅ passedfull logs · dashboard
Pull request closed