From b1e89a45aee2de464c710acc76f19f72fc92f1c6 Mon Sep 17 00:00:00 2001 From: Hou Zhijie Date: Wed, 26 Apr 2023 18:12:19 +0800 Subject: [PATCH v2 1/2] Fix assert failure in logical replication apply worker. The apply worker attempted to release the lock prior to completion of the process initialization, resulting in an assert failure. To resolve the issue, the lock is now released only if the apply worker has finished the initialization phase. --- src/backend/replication/logical/applyparallelworker.c | 4 ++++ src/backend/replication/logical/launcher.c | 6 +++++- src/backend/replication/logical/worker.c | 7 +++++++ src/include/replication/worker_internal.h | 2 ++ 4 files changed, 18 insertions(+), 1 deletion(-) diff --git a/src/backend/replication/logical/applyparallelworker.c b/src/backend/replication/logical/applyparallelworker.c index 4518683779..ee7a18137f 100644 --- a/src/backend/replication/logical/applyparallelworker.c +++ b/src/backend/replication/logical/applyparallelworker.c @@ -873,6 +873,8 @@ ParallelApplyWorkerMain(Datum main_arg) int worker_slot = DatumGetInt32(main_arg); char originname[NAMEDATALEN]; + InitializingApplyWorker = true; + /* Setup signal handling. */ pqsignal(SIGHUP, SignalHandlerForConfigReload); pqsignal(SIGINT, SignalHandlerForShutdownRequest); @@ -940,6 +942,8 @@ ParallelApplyWorkerMain(Datum main_arg) InitializeApplyWorker(); + InitializingApplyWorker = false; + /* Setup replication origin tracking. */ StartTransactionCommand(); ReplicationOriginNameForLogicalRep(MySubscription->oid, InvalidOid, diff --git a/src/backend/replication/logical/launcher.c b/src/backend/replication/logical/launcher.c index 970d170e73..70265b7fd6 100644 --- a/src/backend/replication/logical/launcher.c +++ b/src/backend/replication/logical/launcher.c @@ -797,8 +797,12 @@ logicalrep_worker_onexit(int code, Datum arg) * Session level locks may be acquired outside of a transaction in * parallel apply mode and will not be released when the worker * terminates, so manually release all locks before the worker exits. + * + * However, if the worker is being initialized, there is no need to release + * locks. */ - LockReleaseAll(DEFAULT_LOCKMETHOD, true); + if (!InitializingApplyWorker) + LockReleaseAll(DEFAULT_LOCKMETHOD, true); ApplyLauncherWakeup(); } diff --git a/src/backend/replication/logical/worker.c b/src/backend/replication/logical/worker.c index dbf88c9553..879309b316 100644 --- a/src/backend/replication/logical/worker.c +++ b/src/backend/replication/logical/worker.c @@ -331,6 +331,9 @@ static TransactionId stream_xid = InvalidTransactionId; */ static uint32 parallel_stream_nchanges = 0; +/* Are we initializing a apply worker? */ +bool InitializingApplyWorker = false; + /* * We enable skipping all data modification changes (INSERT, UPDATE, etc.) for * the subscription if the remote transaction's finish LSN matches the subskiplsn. @@ -4526,6 +4529,8 @@ ApplyWorkerMain(Datum main_arg) WalRcvStreamOptions options; int server_version; + InitializingApplyWorker = true; + /* Attach to slot */ logicalrep_worker_attach(worker_slot); @@ -4548,6 +4553,8 @@ ApplyWorkerMain(Datum main_arg) InitializeApplyWorker(); + InitializingApplyWorker = false; + /* Connect to the origin and start the replication. */ elog(DEBUG1, "connecting to publisher using connection string \"%s\"", MySubscription->conninfo); diff --git a/src/include/replication/worker_internal.h b/src/include/replication/worker_internal.h index dce71d2c50..b57eed052f 100644 --- a/src/include/replication/worker_internal.h +++ b/src/include/replication/worker_internal.h @@ -225,6 +225,8 @@ extern PGDLLIMPORT LogicalRepWorker *MyLogicalRepWorker; extern PGDLLIMPORT bool in_remote_transaction; +extern PGDLLIMPORT bool InitializingApplyWorker; + extern void logicalrep_worker_attach(int slot); extern LogicalRepWorker *logicalrep_worker_find(Oid subid, Oid relid, bool only_running); -- 2.30.0.windows.2