Re: Issues with ON CONFLICT UPDATE and REINDEX CONCURRENTLY

Michael Paquier <michael@paquier.xyz>

From: Michael Paquier <michael@paquier.xyz>
To: Álvaro Herrera <alvherre@kurilemu.de>
Cc: Mihail Nikalayeu <mihailnikalayeu@gmail.com>, Noah Misch <noah@leadboat.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2025-11-24T22:15:19Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →
  1. Replace flaky CIC/RI isolation tests with a TAP test

  2. Disable recently added CIC/RI isolation tests

  3. Fix infer_arbiter_index for partitioned tables

  4. Stabilize tests some more

  5. Put back alternative-output expected files

  6. Remove doc and code comments about ON CONFLICT deficiencies

  7. Avoid use of NOTICE to wait for snapshot invalidation

  8. Fix ON CONFLICT with REINDEX CONCURRENTLY and partitions

  9. Fix ON CONFLICT ON CONSTRAINT during REINDEX CONCURRENTLY

  10. Fix new test for CATCACHE_FORCE_RELEASE builds

  11. Improve test case stability

  12. Fix infer_arbiter_index during concurrent index operations

  13. Doc: cover index CONCURRENTLY causing errors in INSERT ... ON CONFLICT.

  14. Fix infer_arbiter_indexes() to not assume resultRelation is 1.

  15. Revert temporal primary keys and foreign keys

On Mon, Nov 24, 2025 at 06:49:47PM +0100, Alvaro Herrera wrote:
> This patch would bring the committed test file up to date with what you
> last submitted.  However, I didn't understand what is the problem with
> the original formulation, and I haven't seen the test fail ... can you
> explain?

Reading through bc32a12e0db2, I am puzzled by the committed result
here:
+#ifdef USE_INJECTION_POINTS
+   if (conflict)
+       INJECTION_POINT("check-exclusion-or-unique-constraint-conflict", NULL);
+   else
+       INJECTION_POINT("check-exclusion-or-unique-constraint-no-conflict", NULL);
+#endif

The "no-conflict" point is used in the isolation test, but not the
other in the "conflict == true" path:
$ git grep check-exclusion-or-unique-constraint-conflict
src/backend/executor/execIndexing.c:
INJECTION_POINT("check-exclusion-or-unique-constraint-conflict", NULL);

If you have no plans for it in the long-term, I'd rather remove it
from the tree, rather than keep it.  Of course, I would keep the
USE_INJECTION_POINTS block to avoid the extra boolean check in
non-USE_INJECTION_POINTS builds.
--
Michael