| From: | Fujii Masao <fujii(at)postgresql(dot)org> |
|---|---|
| To: | pgsql-committers(at)lists(dot)postgresql(dot)org |
| Subject: | pgsql: Fix recovery target test waiting on unavailable WAL |
| Date: | 2026-07-20 18:28:05 |
| Message-ID: | E1wlsij-000000003oZ-0bcX@gemulon.postgresql.org |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-committers |
Fix recovery target test waiting on unavailable WAL
Buildfarm member skink reported a failure in
recovery/003_recovery_targets after commit d5751c33cc3. The newly
added recovery_target_xid set-then-cleared test could time out while
waiting for pg_last_wal_replay_lsn() to reach the expected LSN.
The test recorded lsn6 after calling pg_switch_wal(). As a result,
lsn6 pointed into the next WAL segment, but pg_switch_wal() only
archived the previous one. Since the standby in this test restores WAL
from the archive only, it could not obtain the segment containing
lsn6 and waited indefinitely.
Fix this by recording lsn6 before calling pg_switch_wal(), so the
archived WAL contains the LSN that the standby is waiting for.
Per buildfarm member skink.
Reported-by: Álvaro Herrera <alvherre(at)kurilemu(dot)de>
Author: Fujii Masao <masao(dot)fujii(at)gmail(dot)com>
Discussion: https://postgr.es/m/al5Y2mWffRs1NP34@alvherre.pgsql
Branch
------
master
Details
-------
https://git.postgresql.org/pg/commitdiff/83486b6a65aa48ed18e4e872d5e0fb5b0bfaca41
Modified Files
--------------
src/test/recovery/t/003_recovery_targets.pl | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Masahiko Sawada | 2026-07-21 00:12:37 | pgsql: Add logical decoding status to pg_control_checkpoint(). |
| Previous Message | Robert Haas | 2026-07-20 18:08:24 | pgsql: doc: Granting TRIGGER or REFERENCES on table is dangerous. |