jok-carbonarc commented on PR #2581: URL: https://github.com/apache/age/pull/2581#issuecomment-6439821753
Thanks for the fix. Could #2581 go to the `PG18` branch in the next batch from master, with a 1.8.1 tag? @jrgemignani `PG18` is still at `e43dc1a` (1.8.0), so PG18 users have no release with it. The commit cherry-picks cleanly onto `e43dc1a`. We run 1.8.0 on PostgreSQL 18.6 with `age` in `shared_preload_libraries` and reproduced #2580 with `repro.sh` from the issue, ten fresh servers per row: | build | `rollback` | `quit` (control) | `rollback`, failure = `SET LOCAL statement_timeout` expiring inside `BEGIN`, `CREATE EXTENSION age` done | |---|---|---|---| | 1.8.0 | 10/10 reproduced | 0/10 | 10/10 reproduced | | 1.8.0 + #2581 | 0/10 | 0/10 | 0/10 | The last column is a common client pattern: a per-transaction statement timeout fires and the driver sends `ROLLBACK`. So any PG18 application with timeouts inside explicit transactions can poison PGPROCs, given enough invalidation traffic on the server. #2577 (B-tree strategy numbers) is another master-only fix PG18 users would want in the same batch. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
