MDEV-6403: Temporary tables lost at STOP SLAVE in GTID mode if master has not rotated binlog since restart

The binlog contains specially marked format description events to mark
when a master restart happened (which could have caused temporary
tables to be silently dropped). Such events also cause slave to close
temporary tables.

However, there was a bug that if after this, slave re-connects to the
master in GTID mode, the master can send an old format description
event again. If temporary tables are closed when such event is seen
for the second time, it might drop temporary tables created after that
event, and cause replication failure.

With this patch, the restart flag of the format description event is
cleared by the master when it is sent to the slave in a subsequent
connection, to avoid the errorneous temp table close.

1 files changed, 25 insertions, 0 deletions

diff --git a/sql/rpl_gtid.cc b/sql/rpl_gtid.cc
index d4c7b2c3fbf..f17ece298d3 100644
--- a/sql/rpl_gtid.cc
+++ b/sql/rpl_gtid.cc
@@ -1870,6 +1870,31 @@ slave_connection_state::get_gtid_list(rpl_gtid *gtid_list, uint32 list_size)
 
 
 /*
+  Check if the GTID position has been reached, for mysql_binlog_send().
+
+  The position has not been reached if we have anything in the state, unless
+  it has either the START_ON_EMPTY_DOMAIN flag set (which means it does not
+  belong to this master at all), or the START_OWN_SLAVE_POS (which means that
+  we start on an old position from when the server was a slave with
+  --log-slave-updates=0).
+*/
+bool
+slave_connection_state::is_pos_reached()
+{
+  uint32 i;
+
+  for (i= 0; i < hash.records; ++i)
+  {
+    entry *e= (entry *)my_hash_element(&hash, i);
+    if (!(e->flags & (START_OWN_SLAVE_POS|START_ON_EMPTY_DOMAIN)))
+      return false;
+  }
+
+  return true;
+}
+
+
+/*
   Execute a MASTER_GTID_WAIT().
   The position to wait for is in gtid_str in string form.
   The timeout in microseconds is in timeout_us, zero means no timeout.