| From: | Zsolt Parragi <zsolt(dot)parragi(at)percona(dot)com> |
|---|---|
| To: | Andrey Borodin <x4mmm(at)yandex-team(dot)ru> |
| Cc: | Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com>, pgsql-hackers mailing list <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: SSI: ON CONFLICT DO SELECT takes no predicate lock on the returned row |
| Date: | 2026-10-01 08:28:04 |
| Message-ID: | CAN4CZFPzwmTjbsu_hVu2bFENCssmycf4d1etjOn=mNq7dHwcoQ@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> I found a caller in pg_lake [0] that creates its own EState without
> setting es_snapshot. It pushes an active snapshot, but passes the
> unchanged EState to ExecCheckIndexConstraints(). Your proposed
> InvalidSnapshot guard would cover this too.
Maybe we should open a bug report to pg_lake about this? Even if we
allow for this here, the documentation still requires es_snapshot to
be set.
I attached v2 based on Dean's suggestion, if we want to go that route.
I would still be leaning towards including both code changes instead
(or if we do not want that, adjusting the documentation about this),
but for now this is a simple alternative patch that only relaxes the
original fix when we use es_snapshot.
| Attachment | Content-Type | Size |
|---|---|---|
| v2-0001-Fix-crash-in-logical-replication-apply-worker-at-.patch | application/octet-stream | 4.8 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Zsolt Parragi | 2026-10-01 08:54:05 | Re: REPACK: warn about skipping foreign partitions |
| Previous Message | Álvaro Herrera | 2026-10-01 08:27:49 | Re: Commitfest PG20-2 is now closed |