A new option --replicate-same-server-id to force a slave to execute queries originating from itself

(WL#794). This can be of interest in some recovery-from-backup scenarios, and also when you have
two databases in one mysqld, having a certain similarity and you want one db to be updated when the other is
(some sort of trigger).
Plus small fix for BUG#3568 "MySQL server crashes when built --with-debug and CHANGE MASTER +MASTER_POS_WAIT"


sql/mysqld.cc:
  new option --replicate-same-server-id
sql/slave.cc:
  new option replicate_same_server_id, to force a slave to execute its own queries.
  Small fix for BUG#3568 "MySQL server crashes when built --with-debug and CHANGE MASTER +MASTER_POS_WAIT"
sql/slave.h:
  new option --replicate-same-server-id

mysql-test/r/rpl_server_id1.result

+slave stop;
+drop table if exists t1,t2,t3,t4,t5,t6,t7,t8,t9;
+reset master;
+reset slave;
+drop table if exists t1,t2,t3,t4,t5,t6,t7,t8,t9;
+slave start;
+drop table if exists t1;
+create table t1 (n int);
+reset master;
+stop slave;
+change master to master_port=9307;
+show slave status;
+Master_Host	Master_User	Master_Port	Connect_retry	Master_Log_File	Read_Master_Log_Pos	Relay_Log_File	Relay_Log_Pos	Relay_Master_Log_File	Slave_IO_Running	Slave_SQL_Running	Replicate_do_db	Replicate_ignore_db	Last_errno	Last_error	Skip_counter	Exec_master_log_pos	Relay_log_space
+127.0.0.1	root	SLAVE_PORT	1		4	slave-relay-bin.001	4		No	No			0		0	0	#
+start slave;
+insert into t1 values (1);
+show status like "slave_running";
+Variable_name	Value
+Slave_running	ON
+select * from t1;
+n
+1
+drop table t1;

mysql-test/r/rpl_server_id2.result

+slave stop;
+drop table if exists t1,t2,t3,t4,t5,t6,t7,t8,t9;
+reset master;
+reset slave;
+drop table if exists t1,t2,t3,t4,t5,t6,t7,t8,t9;
+slave start;
+drop table if exists t1;
+create table t1 (n int);
+reset master;
+stop slave;
+change master to master_port=9307;
+show slave status;
+Master_Host	Master_User	Master_Port	Connect_retry	Master_Log_File	Read_Master_Log_Pos	Relay_Log_File	Relay_Log_Pos	Relay_Master_Log_File	Slave_IO_Running	Slave_SQL_Running	Replicate_do_db	Replicate_ignore_db	Last_errno	Last_error	Skip_counter	Exec_master_log_pos	Relay_log_space
+127.0.0.1	root	SLAVE_PORT	1		4	slave-relay-bin.001	4		No	No			0		0	0	#
+start slave;
+insert into t1 values (1);
+select * from t1;
+n
+1
+1
+drop table t1;

mysql-test/t/rpl_server_id1.test

+# This test checks that a slave does not execute queries originating
+# from itself, by default.
+
+source include/master-slave.inc;
+connection slave;
+drop table if exists t1;
+create table t1 (n int);
+reset master;
+# replicate ourselves
+stop slave;
+eval change master to master_port=$SLAVE_MYPORT;
+--replace_result $SLAVE_MYPORT SLAVE_PORT
+--replace_column 18 #
+show slave status;
+start slave;
+insert into t1 values (1);
+# can't MASTER_POS_WAIT(), it does not work in this weird setup
+# (when slave is its own master without --replicate-same-server-id)
+sleep 2; # enough time for the event to be replicated (it should not)
+show status like "slave_running";
+select * from t1;
+drop table t1;

mysql-test/t/rpl_server_id2-slave.opt

+--disable-log-slave-updates --replicate-same-server-id

mysql-test/t/rpl_server_id2.test

+# This test checks that a slave DOES execute queries originating
+# from itself, if running with --replicate-same-server-id.
+
+source include/master-slave.inc;
+connection slave;
+drop table if exists t1;
+create table t1 (n int);
+reset master;
+# replicate ourselves
+stop slave;
+eval change master to master_port=$SLAVE_MYPORT;
+--replace_result $SLAVE_MYPORT SLAVE_PORT
+--replace_column 18 #
+show slave status;
+start slave;
+insert into t1 values (1);
+save_master_pos;
+sync_with_master;
+select * from t1; # check that indeed 2 were inserted
+drop table t1;

sql/mysqld.cc

@@ -2379,6 +2379,15 @@ Warning: you need to use --log-bin to make --log-slave-updates work. \
 Now disabling --log-slave-updates.");
   }
 
+  if (opt_log_slave_updates && replicate_same_server_id)
+  {
+      sql_print_error("\
+Error: using --replicate-same-server-id in conjunction with \
+--log-slave-updates is impossible, it would lead to infinite loops in this \
+server.");
+      unireg_abort(1);
+  }
+
   if (opt_bootstrap)
   {
     int error=bootstrap(stdin);
@@ -3203,7 +3212,7 @@ enum options_mysqld {
   OPT_SKIP_SLAVE_START,        OPT_SKIP_INNOBASE,
   OPT_SAFEMALLOC_MEM_LIMIT,    OPT_REPLICATE_DO_TABLE, 
   OPT_REPLICATE_IGNORE_TABLE,  OPT_REPLICATE_WILD_DO_TABLE, 
-  OPT_REPLICATE_WILD_IGNORE_TABLE, 
+  OPT_REPLICATE_WILD_IGNORE_TABLE, OPT_REPLICATE_SAME_SERVER_ID,
   OPT_DISCONNECT_SLAVE_EVENT_COUNT, 
   OPT_ABORT_SLAVE_EVENT_COUNT,
   OPT_INNODB_DATA_HOME_DIR,
@@ -3613,6 +3622,13 @@ Does nothing yet.",
   {"replicate-rewrite-db", OPT_REPLICATE_REWRITE_DB,
    "Updates to a database with a different name than the original. Example: replicate-rewrite-db=master_db_name->slave_db_name",
    0, 0, 0, GET_STR, REQUIRED_ARG, 0, 0, 0, 0, 0, 0},
+  {"replicate-same-server-id", OPT_REPLICATE_SAME_SERVER_ID,
+   "In replication, if set to 1, do not skip events having our server id. \
+Default value is 0 (to break infinite loops in circular replication). \
+Can't be set to 1 if --log-slave-updates is used.",
+   (gptr*) &replicate_same_server_id,
+   (gptr*) &replicate_same_server_id,
+   0, GET_BOOL, NO_ARG, 0, 0, 0, 0, 0, 0},
   // In replication, we may need to tell the other servers how to connect
   {"report-host", OPT_REPORT_HOST,
    "Hostname or IP of the slave to be reported to to the master during slave registration. Will appear in the output of SHOW SLAVE HOSTS. Leave unset if you do not want the slave to register itself with the master. Note that it is not sufficient for the master to simply read the IP of the slave off the socket once the slave connects. Due to NAT and other routing issues, that IP may not be valid for connecting to the slave from the master or other hosts.",

sql/slave.cc

@@ -38,7 +38,7 @@ HASH replicate_do_table, replicate_ignore_table;
 DYNAMIC_ARRAY replicate_wild_do_table, replicate_wild_ignore_table;
 bool do_table_inited = 0, ignore_table_inited = 0;
 bool wild_do_table_inited = 0, wild_ignore_table_inited = 0;
-bool table_rules_on = 0;
+bool table_rules_on= 0, replicate_same_server_id;
 ulonglong relay_log_space_limit = 0;
 
 /*
@@ -1999,6 +1999,8 @@ int st_relay_log_info::wait_for_pos(THD* thd, String* log_name,
     error= -2; //means improper arguments
     goto err;
   }
+  // Convert 0-3 to 4
+  log_pos= max(log_pos, BIN_LOG_HEADER_SIZE);
   /* p points to '.' */
   log_name_extension= strtoul(++p, &p_end, 10);
   /*
@@ -2019,7 +2021,18 @@ int st_relay_log_info::wait_for_pos(THD* thd, String* log_name,
   {
     bool pos_reached;
     int cmp_result= 0;
-    DBUG_ASSERT(*master_log_name || master_log_pos == 0);
+    /*
+      master_log_name can be "", if we are just after a fresh replication start
+      or after a CHANGE MASTER TO MASTER_HOST/PORT (before we have executed one
+      Rotate event from the master) or (rare) if the user is doing a weird
+      slave setup (see next paragraph).
+      If master_log_name is "", we assume we don't have enough info to do the
+      comparison yet, so we just wait until more data. In this case
+      master_log_pos is always 0 except if somebody (wrongly) sets this slave
+      to be a slave of itself without using --replicate-same-server-id (an
+      unsupported configuration which does nothing), then master_log_pos will
+      grow and master_log_name will stay "".
+    */
     if (*master_log_name)
     {
       char *basename= master_log_name + dirname_length(master_log_name);
@@ -2039,14 +2052,15 @@ int st_relay_log_info::wait_for_pos(THD* thd, String* log_name,
       char *q_end;
       ulong master_log_name_extension= strtoul(q, &q_end, 10);
       if (master_log_name_extension < log_name_extension)
-        cmp_result = -1 ;
+        cmp_result= -1 ;
       else
         cmp_result= (master_log_name_extension > log_name_extension) ? 1 : 0 ;
+
+      pos_reached= ((!cmp_result && master_log_pos >= (ulonglong)log_pos) ||
+                    cmp_result > 0);
+      if (pos_reached || thd->killed)
+        break;
     }
-    pos_reached = ((!cmp_result && master_log_pos >= (ulonglong)log_pos) ||
-                   cmp_result > 0);
-    if (pos_reached || thd->killed)
-      break;
 
     //wait for master update, with optional timeout.
     
@@ -2351,7 +2365,12 @@ static int exec_relay_log_event(THD* thd, RELAY_LOG_INFO* rli)
       log files themselves.
     */
 
-    if (ev->server_id == (uint32) ::server_id ||
+    /*
+      TODO: when this is merged into 4.1, one needs to update queue_event() to
+      add a similar test for replicate_same_server_id, because in 4.1 the I/O
+      thread is also filtering events based on the server id.
+    */
+    if ((ev->server_id == (uint32) ::server_id && !replicate_same_server_id) ||
 	(rli->slave_skip_counter && type_code != ROTATE_EVENT))
     {
       /* TODO: I/O thread should not even log events with the same server id */

sql/slave.h

@@ -418,7 +418,7 @@ extern HASH replicate_do_table, replicate_ignore_table;
 extern DYNAMIC_ARRAY  replicate_wild_do_table, replicate_wild_ignore_table;
 extern bool do_table_inited, ignore_table_inited,
 	    wild_do_table_inited, wild_ignore_table_inited;
-extern bool table_rules_on;
+extern bool table_rules_on, replicate_same_server_id;
 
 extern int disconnect_slave_event_count, abort_slave_event_count ;