Hi, On 2026-04-07 21:05:40 +0800, Xuneng Zhou wrote: > I’ve posted two patches. The first fixes the duplication issue > reported by Andres and is fairly straightforward. The second turned > out to be more complex than expected, and I’m still working through > possible solutions. Feedback or alternative approaches would be very > helpful. > I also spent some time drafting a patch to address the memory ordering > issue and will post it later.
I propose quickly applying a minimal patch like the attached, to get the test performance back to normal. Will do so unless somebody protests within in one CI cycle and one coffee. Greetings, Andres Freund
>From 0a9c10fe36c6b2d08d1f4fbd0825b76bdd389c10 Mon Sep 17 00:00:00 2001 From: Andres Freund <[email protected]> Date: Tue, 7 Apr 2026 09:11:07 -0400 Subject: [PATCH v1] Minimal fix for WAIT FOR ... MODE 'standby_flush' The investigation into the negative test performance impact of 7e8aeb9e483 lead to discovering that there are a few issues with WAIT FOR. This commit is just a minimal fix to prevent hangs in standby_flush mode, due to WAIT FOR ... 'standby_flush' seeing a 0 LSN if a newly started walreceiver does not receive any writes, because the stanby is already caught up. There are several other issues and this is isn't necessarily the best fix. But this way we get the hangs out of the way. Reported-by: Tom Lane <[email protected]> Discussion: https://postgr.es/m/zqbppucpmkeqecfy4s5kscnru4tbk6khp3ozqz6ad2zijz354k@w4bdf4z3wqoz --- src/backend/replication/walreceiver.c | 2 -- src/backend/replication/walreceiverfuncs.c | 1 + 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/src/backend/replication/walreceiver.c b/src/backend/replication/walreceiver.c index a437273cf9a..09fde92bfd7 100644 --- a/src/backend/replication/walreceiver.c +++ b/src/backend/replication/walreceiver.c @@ -242,8 +242,6 @@ WalReceiverMain(const void *startup_data, size_t startup_data_len) SpinLockRelease(&walrcv->mutex); - pg_atomic_write_u64(&WalRcv->writtenUpto, 0); - /* Arrange to clean up at walreceiver exit */ on_shmem_exit(WalRcvDie, PointerGetDatum(&startpointTLI)); diff --git a/src/backend/replication/walreceiverfuncs.c b/src/backend/replication/walreceiverfuncs.c index 4e03e721872..bd5d47be964 100644 --- a/src/backend/replication/walreceiverfuncs.c +++ b/src/backend/replication/walreceiverfuncs.c @@ -321,6 +321,7 @@ RequestXLogStreaming(TimeLineID tli, XLogRecPtr recptr, const char *conninfo, walrcv->flushedUpto = recptr; walrcv->receivedTLI = tli; walrcv->latestChunkStart = recptr; + pg_atomic_write_u64(&walrcv->writtenUpto, recptr); } walrcv->receiveStart = recptr; walrcv->receiveStartTLI = tli; -- 2.53.0.1.gb2826b52eb
