From 8cd687c44b62cda3cc6dec88d414e8e2837dc56b Mon Sep 17 00:00:00 2001 From: Michael Paquier Date: Thu, 11 Jun 2026 17:29:41 +0900 Subject: [PATCH] Fix race with timeline selection in logical decoding during promotion During promotion, there is a window where RecoveryInProgress() returns true but the WAL segments of the old timeline have already been removed. A logical decoding could pick up the old timeline in this window when reading a page, failing with the following error: ERROR: requested WAL segment ... has already been removed This issue does not lead to any data correctness issue, as retrying to decode the data works in follow-up decoding attempts. It impacts availability, though. Other WAL page read callbacks have a similar issue, this commit takes care of what should be the noisiest code path: logical decoding with START_REPLICATION in a WAL sender. A TAP test, based on an injection point waiting in the startup process after the segments have been removed/recycled, is added. This part is backpatched down to v17. This issue has been causing sporadic failures in the buildfarm, and was reproducible manually. This issue happens since logical decoding on standbys exists, down to v16. Reported-by: Alexander Lakhin Author: Bertrand Drouvot Reviewed-by: Hayato Kuroda Reviewed-by: Xuneng Zhou Discussion: https://postgr.es/m/7daef094-abf3-4672-bc23-3df4763b16a3@gmail.com Backpatch-through: 16 --- src/backend/replication/walsender.c | 24 +++++++++++++++++++++++- 1 file changed, 23 insertions(+), 1 deletion(-) diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c index b37f1d19b27..fcf22b7cd14 100644 --- a/src/backend/replication/walsender.c +++ b/src/backend/replication/walsender.c @@ -941,7 +941,29 @@ logical_read_xlog_page(XLogReaderState *state, XLogRecPtr targetPagePtr, int req am_cascading_walsender = RecoveryInProgress(); if (am_cascading_walsender) - GetXLogReplayRecPtr(&currTLI); + { + TimeLineID insertTLI; + + /* + * If the insertion timeline has already been set, use it. + * InsertTimeLineID is set before the WAL segments of the old timeline + * are removed, before SharedRecoveryState switches to + * RECOVERY_STATE_DONE. + * + * There is a window where RecoveryInProgress() still returns true but + * the old timeline's WAL segments have already been removed or + * recycled. Using the WAL insertion timeline avoids attempting to + * read from those removed segments, improving availability, and is a + * safe thing to do as promotion copies the contents in the last + * segment of the old timeline to the first segment of the new + * timeline, up to the switchpoint. + */ + insertTLI = GetWALInsertionTimeLineIfSet(); + if (insertTLI != 0) + currTLI = insertTLI; + else + GetXLogReplayRecPtr(&currTLI); + } else currTLI = GetWALInsertionTimeLine();