Ensure to send a prepare after we detect concurrent abort during decoding.

It is possible that while decoding a prepared transaction, it gets aborted
concurrently via a ROLLBACK PREPARED command. In that case, we were
skipping all the changes and directly sending Rollback Prepared when we
find the same in WAL. However, the downstream has no idea of the GID of
such a transaction. So, ensure to send prepare even when a concurrent
abort is detected.

Author: Ajin Cherian
Reviewed-by: Markus Wanner, Amit Kapila
Discussion: https://postgr.es/m/f82133c6-6055-b400-7922-97dae9f2b50b@enterprisedb.com

diff --git a/doc/src/sgml/logicaldecoding.sgml b/doc/src/sgml/logicaldecoding.sgml
index da23f89ca32c515aba415854925cd666f11e5065..5d049cdc6875a45cbed2842410608ee61a72f8c7 100644 (file)
--- a/doc/src/sgml/logicaldecoding.sgml
+++ b/doc/src/sgml/logicaldecoding.sgml
@@ -545,12 +545,15 @@ CREATE TABLE another_catalog_table(data text) WITH (user_catalog_table = true);
      executed within that transaction. A transaction that is prepared for
      a two-phase commit using <command>PREPARE TRANSACTION</command> will
      also be decoded if the output plugin callbacks needed for decoding
-     them are provided. It is possible that the current transaction which
-     is being decoded is aborted concurrently via a <command>ROLLBACK PREPARED</command>
-     command. In that case, the logical decoding of this transaction will
-     be aborted too. We will skip all the changes of such a transaction once
-     the abort is detected and abort the transaction when we read WAL for
-     <command>ROLLBACK PREPARED</command>.
+     them are provided. It is possible that the current prepared transaction
+     which is being decoded is aborted concurrently via a
+     <command>ROLLBACK PREPARED</command> command. In that case, the logical
+     decoding of this transaction will be aborted too. All the changes of such
+     a transaction are skipped once the abort is detected and the
+     <function>prepare_cb</function> callback is invoked. Thus even in case of
+     a concurrent abort, enough information is provided to the output plugin
+     for it to properly deal with <command>ROLLBACK PREPARED</command> once
+     that is decoded.
     </para>
 
     <note>

diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c
index 127f2c4b168953d93b384ebe788e606fe55cfd08..52d06285a21528dd5fa8b26fc9dffc65ead2dfc5 100644 (file)
--- a/src/backend/replication/logical/reorderbuffer.c
+++ b/src/backend/replication/logical/reorderbuffer.c
@@ -2664,6 +2664,14 @@ ReorderBufferPrepare(ReorderBuffer *rb, TransactionId xid,
 
    ReorderBufferReplay(txn, rb, xid, txn->final_lsn, txn->end_lsn,
                        txn->commit_time, txn->origin_id, txn->origin_lsn);
+
+   /*
+    * We send the prepare for the concurrently aborted xacts so that later
+    * when rollback prepared is decoded and sent, the downstream should be
+    * able to rollback such a xact. See comments atop DecodePrepare.
+    */
+   if (txn->concurrent_abort)
+       rb->prepare(rb, txn, txn->final_lsn);
 }
 
 /*