Many files:

  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3


sql/ha_innodb.cc:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/handler.cc:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/mysqld.cc:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/set_var.cc:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/sql_repl.cc:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/ha_innodb.h:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3
sql/handler.h:
  Semi-synchronous replication for InnoDB type tables; before telling the client that a commit has been processed, wait that the replication thread has returned from my_net_send() where it sends the binlog to the slave; note that TCP/IP, even with the TCP_NODELAY option does not guarantee that the slave has RECEIVED the data - this is just heuristic at the moment; this is useful in failover: in almost all cases, every transaction that has returned from the commit has been sent and processed in the slave, which makes failover to the slave simpler if the master crashes; the code does not work yet as is, because MySQL should call innobase_report_binlog_offset_and_commit() in a commit; we will most probably return that call to 5.0.x, to make InnoDB Hot Backup and group commit to work again; XA code broke them temporarily in 5.0.3

1 files changed, 37 insertions, 0 deletions

diff --git a/sql/sql_repl.cc b/sql/sql_repl.cc
index 634b6ab0995..72470c487a3 100644
--- a/sql/sql_repl.cc
+++ b/sql/sql_repl.cc
@@ -385,6 +385,11 @@ impossible position";
     goto err;
   }
 
+  printf("Binlog file name %s\n", log_file_name);
+
+  if (thd->variables.sync_replication)
+    ha_repl_report_sent_binlog(thd, log_file_name, pos);
+
   /*
     We need to start a packet with something other than 255
     to distinguish it from error
@@ -470,6 +475,10 @@ impossible position";
            my_errno= ER_UNKNOWN_ERROR;
            goto err;
          }
+
+         if (thd->variables.sync_replication)
+           ha_repl_report_sent_binlog(thd, log_file_name, my_b_tell(&log));
+
          /*
            No need to save this event. We are only doing simple reads
            (no real parsing of the events) so we don't need it. And so
@@ -527,6 +536,13 @@ impossible position";
 	my_errno= ER_UNKNOWN_ERROR;
 	goto err;
       }
+
+      printf("Dump loop: %s: Current log position %lu\n", log_file_name,
+                  (ulong)my_b_tell(&log));
+
+      if (thd->variables.sync_replication)
+        ha_repl_report_sent_binlog(thd, log_file_name, my_b_tell(&log));
+
       DBUG_PRINT("info", ("log event code %d",
 			  (*packet)[LOG_EVENT_OFFSET+1] ));
       if ((*packet)[LOG_EVENT_OFFSET+1] == LOAD_EVENT)
@@ -640,6 +656,12 @@ impossible position";
 	    goto err;
 	  }
 
+          printf("Second loop: %s: Current log position %lu\n", log_file_name,
+                  (ulong)my_b_tell(&log));
+
+          if (thd->variables.sync_replication)
+            ha_repl_report_sent_binlog(thd, log_file_name, my_b_tell(&log));
+
 	  if ((*packet)[LOG_EVENT_OFFSET+1] == LOAD_EVENT)
 	  {
 	    if (send_file(thd))
@@ -704,12 +726,22 @@ impossible position";
 	my_errno= ER_MASTER_FATAL_ERROR_READING_BINLOG;
 	goto err;
       }
+
+      if (thd->variables.sync_replication)
+        ha_repl_report_sent_binlog(thd, log_file_name, 0);
+
+      printf("Binlog file name of a new binlog %s\n", log_file_name);
+
       packet->length(0);
       packet->append('\0');
     }
   }
 
 end:
+  printf("Ending replication\n");
+  if (thd->variables.sync_replication)
+    ha_repl_report_replication_stop(thd);
+
   end_io_cache(&log);
   (void)my_close(file, MYF(MY_WME));
 
@@ -721,6 +753,11 @@ end:
   DBUG_VOID_RETURN;
 
 err:
+  if (thd->variables.sync_replication)
+    ha_repl_report_replication_stop(thd);
+
+  printf("Ending replication in error %s\n", errmsg);
+
   thd->proc_info = "Waiting to finalize termination";
   end_io_cache(&log);
   /*