| commit | cbfbda78413a5b2f4807e029407dcc98a0e63162 | |
| author | Heikki Linnakangas <heikki.linnakangas@iki.fi> | |
| Thu, 27 Jun 2024 18:06:32 +0000 (27 21:06 +0300) | ||
| committer | Heikki Linnakangas <heikki.linnakangas@iki.fi> | |
| Thu, 27 Jun 2024 18:09:58 +0000 (27 21:09 +0300) | ||
| tree | e09068019ad2bf82bb6d93aab267385dd24bef42 | treesnapshot (tar.gz zip) |
| parent | ecbf6ac51df27275fb0db493bf163ef98ac00c6a | commitdiff |
Fix MVCC bug with prepared xact with subxacts on standby
We did not recover the subtransaction IDs of prepared transactions
when starting a hot standby from a shutdown checkpoint. As a result,
such subtransactions were considered as aborted, rather than
in-progress. That would lead to hint bits being set incorrectly, and
the subtransactions suddenly becoming visible to old snapshots when
the prepared transaction was committed.
To fix, update pg_subtrans with prepared transactions's subxids when
starting hot standby from a shutdown checkpoint. The snapshots taken
from that state need to be marked as "suboverflowed", so that we also
check the pg_subtrans.
Backport to all supported versions.
Discussion: https://www.postgresql.org/message-id/6b852e98-2d49-4ca1-9e95-db419a2696e0@iki.fi
We did not recover the subtransaction IDs of prepared transactions
when starting a hot standby from a shutdown checkpoint. As a result,
such subtransactions were considered as aborted, rather than
in-progress. That would lead to hint bits being set incorrectly, and
the subtransactions suddenly becoming visible to old snapshots when
the prepared transaction was committed.
To fix, update pg_subtrans with prepared transactions's subxids when
starting hot standby from a shutdown checkpoint. The snapshots taken
from that state need to be marked as "suboverflowed", so that we also
check the pg_subtrans.
Backport to all supported versions.
Discussion: https://www.postgresql.org/message-id/6b852e98-2d49-4ca1-9e95-db419a2696e0@iki.fi