From 632dd304c7d1915d46a1b7e1fcca7821cae6e9d6 Mon Sep 17 00:00:00 2001
From: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
Date: Thu, 11 Jul 2024 06:55:45 -0600
Subject: [PATCH] MDEV-34554: rpl_change_master_demote sporadically fails on
 buildbot

MDEV-34274 did not fix the test failure. The test has a START SLAVE
UNTIL condition, where we can't use sync_with_master_gtid.inc,
wait_for_slave_to_start.inc, or wait_for_slave_to_stop.inc because
our MTR connection thread races with the start/stop of the SQL/IO
threads. So instead, for slave start, we prove the threads started
by waiting for the connection count to increase by 2; and for slave
stop, we wait for the processlist count to return to its pre start
slave number.
---
 .../rpl/r/rpl_change_master_demote.result     |  3 +--
 .../suite/rpl/t/rpl_change_master_demote.test | 21 ++++++++++++-------
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/mysql-test/suite/rpl/r/rpl_change_master_demote.result b/mysql-test/suite/rpl/r/rpl_change_master_demote.result
index dc6e5c2ffee..5924b833709 100644
--- a/mysql-test/suite/rpl/r/rpl_change_master_demote.result
+++ b/mysql-test/suite/rpl/r/rpl_change_master_demote.result
@@ -501,8 +501,7 @@ Warnings:
 Note	1278	It is recommended to use --skip-slave-start when doing step-by-step replication with START SLAVE UNTIL; otherwise, you will get problems if you get an unexpected slave's mariadbd restart
 # Slave needs time to start and stop automatically
 # Waiting for both SQL and IO threads to have started..
-# Waiting for SQL thread to be killed..
-# Waiting for IO thread to be killed..
+# Waiting for Slave SQL and IO threads to be killed..
 # Validating neither SQL nor IO threads are running..
 # ..success
 # Clean slave state of master
diff --git a/mysql-test/suite/rpl/t/rpl_change_master_demote.test b/mysql-test/suite/rpl/t/rpl_change_master_demote.test
index 9754b03f1cc..255e9e34b9c 100644
--- a/mysql-test/suite/rpl/t/rpl_change_master_demote.test
+++ b/mysql-test/suite/rpl/t/rpl_change_master_demote.test
@@ -277,36 +277,43 @@ SELECT VARIABLE_NAME, GLOBAL_VALUE FROM INFORMATION_SCHEMA.SYSTEM_VARIABLES WHER
 --echo # position pointing to a previous event (because
 --echo # master_demote_to_slave=1 merges gtid_binlog_pos into gtid_slave_pos).
 
---let $pre_start_slave_thread_count= query_get_value(SHOW STATUS LIKE 'Connections', Value, 1)
+# Note that we can't use sync_with_master_gtid.inc,
+# wait_for_slave_to_start.inc, or wait_for_slave_to_stop.inc because our MTR
+# connection thread races with the start/stop of the SQL/IO threads. So
+# instead, for slave start, we prove the threads started by waiting for the
+# connection count to increase by 2; and for slave stop, we wait for the
+# processlist count to return to its pre start slave number.
+
+--let $pre_start_slave_conn_count= query_get_value(SHOW STATUS LIKE 'Connections', Value, 1)
+--let $pre_start_slave_process_count= `SELECT count(*) from information_schema.PROCESSLIST`
 
 --replace_result $ssu_middle_binlog_pos ssu_middle_binlog_pos
 eval START SLAVE UNTIL master_gtid_pos="$ssu_middle_binlog_pos";
 
 --echo # Slave needs time to start and stop automatically
 --echo # Waiting for both SQL and IO threads to have started..
---let $expected_cons_after_start_slave= `SELECT ($pre_start_slave_thread_count + 2)`
+--let $expected_cons_after_start_slave= `SELECT ($pre_start_slave_conn_count + 2)`
 --let $status_var= Connections
 --let $status_var_value= $expected_cons_after_start_slave
 --let $status_var_comparsion= >=
 --source include/wait_for_status_var.inc
 --let $status_var_comparsion=
 
---echo # Waiting for SQL thread to be killed..
---let $wait_condition= SELECT count(*)=0 from information_schema.PROCESSLIST where COMMAND="Slave_SQL"
---source include/wait_condition.inc
---echo # Waiting for IO thread to be killed..
---let $wait_condition= SELECT count(*)=0 from information_schema.PROCESSLIST where COMMAND="Slave_IO"
+--echo # Waiting for Slave SQL and IO threads to be killed..
+--let $wait_condition= SELECT count(*)=$pre_start_slave_process_count from information_schema.PROCESSLIST
 --source include/wait_condition.inc
 
 --echo # Validating neither SQL nor IO threads are running..
 --let $io_state= query_get_value("SHOW SLAVE STATUS", Slave_IO_State, 1)
 if (`SELECT strcmp("$io_state","") != 0`)
 {
+    --echo # Slave_IO_State is "$io_state" but should be empty
     die "IO thread should not be running after START SLAVE UNTIL master_gtid_pos using a pre-existing GTID";
 }
 --let $sql_state= query_get_value("SHOW SLAVE STATUS", Slave_SQL_Running_State, 1)
 if (`SELECT strcmp("$sql_state","") != 0`)
 {
+    --echo # Slave_SQL_Running_State is "$sql_state" but should be empty
     die "SQL thread should not be running after START SLAVE UNTIL master_gtid_pos using a pre-existing GTID";
 }
 --echo # ..success