Manual merge from 5.0-rpl, of fixes for:

1) BUG#25507 "multi-row insert delayed + auto increment causes duplicate key entries on slave" (two concurrrent connections doing multi-row INSERT DELAYED to insert into an auto_increment column, caused replication slave to stop with "duplicate key error" (and binlog was wrong), and BUG#26116 "If multi-row INSERT DELAYED has errors, statement-based binlogging breaks" (the binlog was not accounting for all rows inserted, or slave could stop). The fix is that: in statement-based binlogging, a multi-row INSERT DELAYED is silently converted to a non-delayed INSERT. This is supposed to not affect many 5.1 users as in 5.1, the default binlog format is "mixed", which does not have the bug (the bug is only with binlog_format=STATEMENT). We should document how the system delayed_insert thread decides of its binlog format (which is not modified by this patch): this decision is taken when the thread is created and holds until it is terminated (is not affected by any later change via SET GLOBAL BINLOG_FORMAT). It is also not affected by the binlog format of the connection which issues INSERT DELAYED (this binlog format does not affect how the row will be binlogged). If one wants to change the binlog format of its server with SET GLOBAL BINLOG_FORMAT, it should do FLUSH TABLES to be sure all delayed_insert threads terminate and thus new threads are created, taking into account the new format. 2) BUG#24432 "INSERT... ON DUPLICATE KEY UPDATE skips auto_increment values". When in an INSERT ON DUPLICATE KEY UPDATE, using an autoincrement column, we inserted some autogenerated values and also updated some rows, some autogenerated values were not used (for example, even if 10 was the largest autoinc value in the table at the start of the statement, 12 could be the first autogenerated value inserted by the statement, instead of 11). One autogenerated value was lost per updated row. Led to exhausting the range of the autoincrement column faster. Bug introduced by fix of BUG#20188; present since 5.0.24 and 5.1.12. This bug breaks replication from a pre-5.0.24/pre-5.1.12 master. But the present bugfix, as it makes INSERT ON DUP KEY UPDATE behave like pre-5.0.24/pre-5.1.12, breaks replication from a [5.0.24,5.0.34]/[5.1.12,5.1.15] master to a fixed (5.0.36/5.1.16) slave! To warn users against this when they upgrade their slave, as agreed with the support team, we add code for a fixed slave to detect that it is connected to a buggy master in a situation (INSERT ON DUP KEY UPDATE into autoinc column) likely to break replication, in which case it cannot replicate so stops and prints a message to the slave's error log and to SHOW SLAVE STATUS. For 5.0.36->[5.0.24,5.0.34] replication or 5.1.16->[5.1.12,5.1.15] replication we cannot warn as master does not know the slave's version (but we always recommended to users to have slave at least as new as master). As agreed with support, I have asked for an alert to be put into the MySQL Network Monitoring and Advisory Service. 3) note that I'll re-enable rpl_insert_id as soon as 5.1-rpl gets the changes from the main 5.1.

Manual merge from 5.0-rpl, of fixes for:
1) BUG#25507 "multi-row insert delayed + auto increment causes duplicate key entries on slave" (two concurrrent connections doing multi-row INSERT DELAYED to insert into an auto_increment column, caused replication slave to stop with "duplicate key error" (and binlog was wrong), and BUG#26116 "If multi-row INSERT DELAYED has errors, statement-based binlogging breaks" (the binlog was not accounting for all rows inserted, or slave could stop). The fix is that: in statement-based binlogging, a multi-row INSERT DELAYED is silently converted to a non-delayed INSERT. This is supposed to not affect many 5.1 users as in 5.1, the default binlog format is "mixed", which does not have the bug (the bug is only with binlog_format=STATEMENT). We should document how the system delayed_insert thread decides of its binlog format (which is not modified by this patch): this decision is taken when the thread is created and holds until it is terminated (is not affected by any later change via SET GLOBAL BINLOG_FORMAT). It is also not affected by the binlog format of the connection which issues INSERT DELAYED (this binlog format does not affect how the row will be binlogged). If one wants to change the binlog format of its server with SET GLOBAL BINLOG_FORMAT, it should do FLUSH TABLES to be sure all delayed_insert threads terminate and thus new threads are created, taking into account the new format. 2) BUG#24432 "INSERT... ON DUPLICATE KEY UPDATE skips auto_increment values". When in an INSERT ON DUPLICATE KEY UPDATE, using an autoincrement column, we inserted some autogenerated values and also updated some rows, some autogenerated values were not used (for example, even if 10 was the largest autoinc value in the table at the start of the statement, 12 could be the first autogenerated value inserted by the statement, instead of 11). One autogenerated value was lost per updated row. Led to exhausting the range of the autoincrement column faster. Bug introduced by fix of BUG#20188; present since 5.0.24 and 5.1.12. This bug breaks replication from a pre-5.0.24/pre-5.1.12 master. But the present bugfix, as it makes INSERT ON DUP KEY UPDATE behave like pre-5.0.24/pre-5.1.12, breaks replication from a [5.0.24,5.0.34]/[5.1.12,5.1.15] master to a fixed (5.0.36/5.1.16) slave! To warn users against this when they upgrade their slave, as agreed with the support team, we add code for a fixed slave to detect that it is connected to a buggy master in a situation (INSERT ON DUP KEY UPDATE into autoinc column) likely to break replication, in which case it cannot replicate so stops and prints a message to the slave's error log and to SHOW SLAVE STATUS. For 5.0.36->[5.0.24,5.0.34] replication or 5.1.16->[5.1.12,5.1.15] replication we cannot warn as master does not know the slave's version (but we always recommended to users to have slave at least as new as master). As agreed with support, I have asked for an alert to be put into the MySQL Network Monitoring and Advisory Service. 3) note that I'll re-enable rpl_insert_id as soon as 5.1-rpl gets the changes from the main 5.1.
39de08fd · guilhem@gbichot3.local · 03dfe4ab · 39de08fd · 39de08fd · 39de08fd
Commit 39de08fd authored Feb 15, 2007 by guilhem@gbichot3.local
12 changed files
--- a/mysql-test/t/rpl_insert_delayed.test
+++ b/mysql-test/t/rpl_insert_delayed.test
--source include/master-slave.inc
--source include/not_embedded.inc
--source include/not_windows.inc
+# The two bugs below (BUG#25507 and BUG#26116) existed only in
+# statement-based binlogging; we test that now they are fixed;
+# we also test that mixed and row-based binlogging work too,
+# for completeness.

 connection master;
+--disable_warnings
+CREATE SCHEMA IF NOT EXISTS mysqlslap;
+USE mysqlslap;
+--enable_warnings

-let $binlog_format_statement=1;
+select @@global.binlog_format;
+
+#
+# BUG#25507 "multi-row insert delayed + auto increment causes
+# duplicate key entries on slave";
+# happened only in statement-based binlogging.
+#

 CREATE TABLE t1 (id INT primary key auto_increment, name VARCHAR(64));
+let $query = "INSERT DELAYED INTO t1 VALUES (null, 'Dr. No'), (null, 'From Russia With Love'), (null, 'Goldfinger'), (null, 'Thunderball'), (null, 'You Only Live Twice')";
+--exec $MYSQL_SLAP --silent --concurrency=5 --iterations=200 --query=$query --delimiter=";"

+FLUSH TABLE t1; # another way to be sure INSERT DELAYED has inserted
+SELECT COUNT(*) FROM t1;
+# when bug existed slave failed below ("duplicate key" error at random INSERT)
 sync_slave_with_master;
+use mysqlslap;
+SELECT COUNT(*) FROM t1;

 #
 # BUG#26116 "If multi-row INSERT DELAYED has errors,
@@ -62,6 +80,7 @@ select * from t1;

 # clean up
 connection master;
-drop table t1;
+USE test;
+DROP SCHEMA mysqlslap;
 sync_slave_with_master;
 connection master;
--- a/mysql-test/extra/rpl_tests/rpl_insert_id.test
+++ b/mysql-test/extra/rpl_tests/rpl_insert_id.test
@@ -209,7 +209,7 @@ connection master;

 drop function bug15728;
 drop function bug15728_insert;
-drop table t1;
+drop table t1,t2;
 drop procedure foo;

 # test of BUG#20188 REPLACE or ON DUPLICATE KEY UPDATE in
@@ -276,6 +276,59 @@ connection master;
 drop table t1;
 sync_slave_with_master;

+#
+# BUG#24432 "INSERT... ON DUPLICATE KEY UPDATE skips auto_increment values"
+#
+
+connection master;
+# testcase with INSERT VALUES
+CREATE TABLE t1 (a INT NOT NULL PRIMARY KEY AUTO_INCREMENT, b INT,
+UNIQUE(b));
+INSERT INTO t1(b) VALUES(1),(1),(2) ON DUPLICATE KEY UPDATE t1.b=10;
+SELECT * FROM t1;
+sync_slave_with_master;
+SELECT * FROM t1;
+connection master;
+drop table t1;
+
+# tescase with INSERT SELECT
+CREATE TABLE t1 (
+  id bigint(20) unsigned NOT NULL auto_increment,
+  field_1 int(10) unsigned NOT NULL,
+  field_2 varchar(255) NOT NULL,
+  field_3 varchar(255) NOT NULL,
+  PRIMARY KEY (id),
+  UNIQUE KEY field_1 (field_1, field_2)
+);
+CREATE TABLE t2 (
+  field_a int(10) unsigned NOT NULL,
+  field_b varchar(255) NOT NULL,
+  field_c varchar(255) NOT NULL
+);
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (1, 'a', '1a');
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (2, 'b', '2b');
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (3, 'c', '3c');
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (4, 'd', '4d');
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (5, 'e', '5e');
+# Updating table t1 based on values from table t2
+INSERT INTO t1 (field_1, field_2, field_3)
+SELECT t2.field_a, t2.field_b, t2.field_c
+FROM t2
+ON DUPLICATE KEY UPDATE
+t1.field_3 = t2.field_c;
+# Inserting new record into t2
+INSERT INTO t2 (field_a, field_b, field_c) VALUES (6, 'f', '6f');
+# Updating t1 again
+INSERT INTO t1 (field_1, field_2, field_3)
+SELECT t2.field_a, t2.field_b, t2.field_c
+FROM t2
+ON DUPLICATE KEY UPDATE
+t1.field_3 = t2.field_c;
+SELECT * FROM t1;
+sync_slave_with_master;
+SELECT * FROM t1;
+connection master;
+drop table t1, t2;

 #
 # BUG#20339: stored procedure using LAST_INSERT_ID() does not

--- a/mysql-test/r/rpl_insert_id.result
+++ b/mysql-test/r/rpl_insert_id.result
@@ -196,7 +196,7 @@ id	last_id
 3	5
 drop function bug15728;
 drop function bug15728_insert;
-drop table t1;
+drop table t1,t2;
 drop procedure foo;
 create table t1 (n int primary key auto_increment not null,
 b int, unique(b));

--- a/mysql-test/r/rpl_known_bugs_detection.result
+++ b/mysql-test/r/rpl_known_bugs_detection.result
@@ -33,7 +33,7 @@ Replicate_Wild_Ignore_Table
 Last_Errno	1105
 Last_Error	Error 'master may suffer from http://bugs.mysql.com/bug.php?id=24432 so slave stops; check error log on slave for more info' on query. Default database: 'test'. Query: 'INSERT INTO t1(b) VALUES(1),(1),(2) ON DUPLICATE KEY UPDATE t1.b=10'
 Skip_Counter	0
-Exec_Master_Log_Pos	238
+Exec_Master_Log_Pos	242
 Relay_Log_Space	#
 Until_Condition	None
 Until_Log_File	
@@ -115,7 +115,7 @@ FROM t2
 ON DUPLICATE KEY UPDATE
 t1.field_3 = t2.field_c'
 Skip_Counter	0
-Exec_Master_Log_Pos	1270
+Exec_Master_Log_Pos	1274
 Relay_Log_Space	#
 Until_Condition	None
 Until_Log_File	

--- a/mysql-test/r/rpl_insert_delayed.result
+++ b/mysql-test/r/rpl_insert_delayed.result
@@ -4,21 +4,36 @@ reset master;
 reset slave;
 drop table if exists t1,t2,t3,t4,t5,t6,t7,t8,t9;
 start slave;
+set @old_global_binlog_format = @@global.binlog_format;
+set @@global.binlog_format = row;
+CREATE SCHEMA IF NOT EXISTS mysqlslap;
+USE mysqlslap;
+select @@global.binlog_format;
+@@global.binlog_format
+ROW
 CREATE TABLE t1 (id INT primary key auto_increment, name VARCHAR(64));
+FLUSH TABLE t1;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
+use mysqlslap;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
 truncate table t1;
 insert delayed into t1 values(10, "my name");
 insert delayed into t1 values(10, "is Bond"), (20, "James Bond");
-ERROR 23000: Duplicate entry '10' for key 1
 flush table t1;
 select * from t1;
 id	name
 10	my name
+20	James Bond
 select * from t1;
 id	name
 10	my name
+20	James Bond
 delete from t1 where id!=10;
 insert delayed into t1 values(20, "is Bond"), (10, "James Bond");
-ERROR 23000: Duplicate entry '10' for key 1
 flush table t1;
 select * from t1;
 id	name
@@ -28,4 +43,6 @@ select * from t1;
 id	name
 10	my name
 20	is Bond
-drop table t1;
+USE test;
+DROP SCHEMA mysqlslap;
+set @@global.binlog_format = @old_global_binlog_format;
--- a/mysql-test/r/rpl_stm_insert_delayed.result
+++ b/mysql-test/r/rpl_stm_insert_delayed.result
+stop slave;
+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;
+start slave;
+set @old_global_binlog_format = @@global.binlog_format;
+set @@global.binlog_format = statement;
+CREATE SCHEMA IF NOT EXISTS mysqlslap;
+USE mysqlslap;
+select @@global.binlog_format;
+@@global.binlog_format
+STATEMENT
+CREATE TABLE t1 (id INT primary key auto_increment, name VARCHAR(64));
+FLUSH TABLE t1;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
+use mysqlslap;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
+truncate table t1;
+insert delayed into t1 values(10, "my name");
+insert delayed into t1 values(10, "is Bond"), (20, "James Bond");
+ERROR 23000: Duplicate entry '10' for key 'PRIMARY'
+flush table t1;
+select * from t1;
+id	name
+10	my name
+select * from t1;
+id	name
+10	my name
+delete from t1 where id!=10;
+insert delayed into t1 values(20, "is Bond"), (10, "James Bond");
+ERROR 23000: Duplicate entry '10' for key 'PRIMARY'
+flush table t1;
+select * from t1;
+id	name
+10	my name
+20	is Bond
+select * from t1;
+id	name
+10	my name
+20	is Bond
+USE test;
+DROP SCHEMA mysqlslap;
+set @@global.binlog_format = mixed;
+CREATE SCHEMA IF NOT EXISTS mysqlslap;
+USE mysqlslap;
+select @@global.binlog_format;
+@@global.binlog_format
+MIXED
+CREATE TABLE t1 (id INT primary key auto_increment, name VARCHAR(64));
+FLUSH TABLE t1;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
+use mysqlslap;
+SELECT COUNT(*) FROM t1;
+COUNT(*)
+5000
+truncate table t1;
+insert delayed into t1 values(10, "my name");
+insert delayed into t1 values(10, "is Bond"), (20, "James Bond");
+flush table t1;
+select * from t1;
+id	name
+10	my name
+20	James Bond
+select * from t1;
+id	name
+10	my name
+20	James Bond
+delete from t1 where id!=10;
+insert delayed into t1 values(20, "is Bond"), (10, "James Bond");
+flush table t1;
+select * from t1;
+id	name
+10	my name
+20	is Bond
+select * from t1;
+id	name
+10	my name
+20	is Bond
+USE test;
+DROP SCHEMA mysqlslap;
+set @@global.binlog_format = @old_global_binlog_format;
--- a/mysql-test/t/disabled.def
+++ b/mysql-test/t/disabled.def
@@ -38,4 +38,5 @@ synchronization          : Bug#24529  	Test 'synchronization' fails on Mac pushb

 flush2                   : Bug#24805 Pushbuild can't handle test with --disable-log-bin
 mysql_upgrade            : Bug#25074 mysql_upgrade gives inconsisten results
+rpl_insert_id            : Guilhem will enable as soon as replication team tree has merged with main tree and so got version number 5.1.16

--- a/mysql-test/t/rpl_known_bugs_detection.test
+++ b/mysql-test/t/rpl_known_bugs_detection.test
@@ -6,6 +6,9 @@
 source include/have_debug.inc;
 source include/master-slave.inc;

+# Currently only statement-based-specific bugs are here
+-- source include/have_binlog_format_mixed_or_statement.inc
+
 #
 # This is to test that slave properly detects if
 # master may suffer from:

--- a/mysql-test/t/rpl_row_insert_delayed.test
+++ b/mysql-test/t/rpl_row_insert_delayed.test
+--source include/have_binlog_format_row.inc
+--source include/master-slave.inc
+--source include/not_embedded.inc
+--source include/not_windows.inc
+
+connection master;
+set @old_global_binlog_format = @@global.binlog_format;
+
+let $binlog_format_statement=0;
+set @@global.binlog_format = row;
+--source extra/rpl_tests/rpl_insert_delayed.test
+
+connection master;
+set @@global.binlog_format = @old_global_binlog_format;
--- a/mysql-test/t/rpl_stm_insert_delayed.test
+++ b/mysql-test/t/rpl_stm_insert_delayed.test
+# we run first in statement-based then in mixed binlogging
+
+--source include/have_binlog_format_mixed_or_statement.inc
+--source include/master-slave.inc
+--source include/not_embedded.inc
+--source include/not_windows.inc
+
+connection master;
+set @old_global_binlog_format = @@global.binlog_format;
+
+let $binlog_format_statement=1;
+set @@global.binlog_format = statement;
+--source extra/rpl_tests/rpl_insert_delayed.test
+
+let $binlog_format_statement=0;
+set @@global.binlog_format = mixed;
+--source extra/rpl_tests/rpl_insert_delayed.test
+
+connection master;
+set @@global.binlog_format = @old_global_binlog_format;
--- a/sql/slave.cc
+++ b/sql/slave.cc
@@ -3684,8 +3684,13 @@ bool rpl_master_has_bug(RELAY_LOG_INFO *rli, uint bug_id)
        (memcmp(introduced_in, master_ver, 3) <= 0) &&
        (memcmp(fixed_in,      master_ver, 3) >  0))
    {
+      // a short message for SHOW SLAVE STATUS (message length constraints)
+      my_printf_error(ER_UNKNOWN_ERROR, "master may suffer from"
+                      " http://bugs.mysql.com/bug.php?id=%u"
+                      " so slave stops; check error log on slave"
+                      " for more info", MYF(0), bug_id);
      // a verbose message for the error log
-      slave_print_error(rli, ER_UNKNOWN_ERROR,
+      slave_print_msg(ERROR_LEVEL, rli, ER_UNKNOWN_ERROR,
                      "According to the master's version ('%s'),"
                      " it is probable that master suffers from this bug:"
                      " http://bugs.mysql.com/bug.php?id=%u"
@@ -3704,11 +3709,6 @@ bool rpl_master_has_bug(RELAY_LOG_INFO *rli, uint bug_id)
                      rli->relay_log.description_event_for_exec->server_version,
                      bug_id,
                      fixed_in[0], fixed_in[1], fixed_in[2]);
-      // a short message for SHOW SLAVE STATUS (message length constraints)
-      my_printf_error(ER_UNKNOWN_ERROR, "master may suffer from"
-                      " http://bugs.mysql.com/bug.php?id=%u"
-                      " so slave stops; check error log on slave"
-                      " for more info", MYF(0), bug_id);
      return TRUE;
    }
  }

--- a/sql/sql_insert.cc
+++ b/sql/sql_insert.cc
@@ -59,6 +59,7 @@
 #include "sql_trigger.h"
 #include "sql_select.h"
 #include "sql_show.h"
+#include "slave.h"

 static int check_null_fields(THD *thd,TABLE *entry);
 #ifndef EMBEDDED_LIBRARY
@@ -341,6 +342,36 @@ bool mysql_insert(THD *thd,TABLE_LIST *table_list,
      (duplic == DUP_UPDATE))
    lock_type=TL_WRITE;
 #endif
+  if ((lock_type == TL_WRITE_DELAYED) &&
+      (global_system_variables.binlog_format == BINLOG_FORMAT_STMT) &&
+      log_on && mysql_bin_log.is_open() &&
+      (values_list.elements > 1))
+  {
+    /*
+      Statement-based binary logging does not work in this case, because:
+      a) two concurrent statements may have their rows intermixed in the
+      queue, leading to autoincrement replication problems on slave (because
+      the values generated used for one statement don't depend only on the
+      value generated for the first row of this statement, so are not
+      replicable)
+      b) if first row of the statement has an error the full statement is
+      not binlogged, while next rows of the statement may be inserted.
+      c) if first row succeeds, statement is binlogged immediately with a
+      zero error code (i.e. "no error"), if then second row fails, query
+      will fail on slave too and slave will stop (wrongly believing that the
+      master got no error).
+      So we fallback to non-delayed INSERT.
+      Note that to be fully correct, we should test the "binlog format which
+      the delayed thread is going to use for this row". But in the common case
+      where the global binlog format is not changed and the session binlog
+      format may be changed, that is equal to the global binlog format.
+      We test it without mutex for speed reasons (condition rarely true), and
+      in the common case (global not changed) it is as good as without mutex;
+      if global value is changed, anyway there is uncertainty as the delayed
+      thread may be old and use the before-the-change value.
+    */
+    lock_type= TL_WRITE;
+  }
  table_list->lock_type= lock_type;

 #ifndef EMBEDDED_LIBRARY
@@ -454,6 +485,14 @@ bool mysql_insert(THD *thd,TABLE_LIST *table_list,
  thd->cuted_fields = 0L;
  table->next_number_field=table->found_next_number_field;

+#ifdef HAVE_REPLICATION
+  if (thd->slave_thread &&
+      (info.handle_duplicates == DUP_UPDATE) &&
+      (table->next_number_field != NULL) &&
+      rpl_master_has_bug(&active_mi->rli, 24432))
+    goto abort;
+#endif
+
  error=0;
  thd->proc_info="update";
  if (duplic != DUP_ERROR || ignore)
@@ -1146,13 +1185,13 @@ int write_record(THD *thd, TABLE *table,COPY_INFO *info)
        if (res == VIEW_CHECK_ERROR)
          goto before_trg_err;

+        table->file->restore_auto_increment(prev_insert_id);
        if ((error=table->file->ha_update_row(table->record[1],
                                              table->record[0])))
 	{
          if (info->ignore &&
              !table->file->is_fatal_error(error, HA_CHECK_DUP_KEY))
          {
-            table->file->restore_auto_increment(prev_insert_id);
            goto ok_or_after_trg_err;
          }
          goto err;
@@ -2489,6 +2528,15 @@ select_insert::prepare(List<Item> &values, SELECT_LEX_UNIT *u)
  }
  restore_record(table,s->default_values);		// Get empty record
  table->next_number_field=table->found_next_number_field;
+
+#ifdef HAVE_REPLICATION
+  if (thd->slave_thread &&
+      (info.handle_duplicates == DUP_UPDATE) &&
+      (table->next_number_field != NULL) &&
+      rpl_master_has_bug(&active_mi->rli, 24432))
+    DBUG_RETURN(1);
+#endif
+
  thd->cuted_fields=0;
  if (info.ignore || info.handle_duplicates != DUP_ERROR)
    table->file->extra(HA_EXTRA_IGNORE_DUP_KEY);