Our custom app went through an install on 2021-07-29 where we dumped the master DB (with master info/pos included), imported it into the new slave, and proceeded to run with replication - this started in line 87 of the attached file.

On 2021-08-04, we upgraded our custom app (does NOT upgrade MariaDB) which runs the following commands between the times shown:

2021-08-04 10:01:39

stop slave;
CHANGE MASTER TO MASTER_CONNECT_RETRY = 1, MASTER_HEARTBEAT_PERIOD = 90, MASTER_USER = 'mvp_repl_slave', MASTER_PASSWORD = '####';
start slave;
DELETE FROM user WHERE !((User='root' AND Host='localhost') OR (User='mariadb.sys' AND Host='localhost'));
FLUSH PRIVILEGES;
GRANT SELECT, INSERT, UPDATE, DELETE, EXECUTE, CREATE, DROP, CREATE VIEW, SHOW VIEW, FILE, SUPER, REPLICATION CLIENT ON . to mvp_local@'localhost' IDENTIFIED BY '####';
GRANT SELECT, EXECUTE, SUPER, REPLICATION CLIENT ON . to mvp_peer@'192.168.2.2' IDENTIFIED BY '####';
GRANT SELECT, EXECUTE, SUPER, REPLICATION CLIENT ON . to mvp_peer@'192.168.2.3' IDENTIFIED BY '####';
FLUSH PRIVILEGES;

2021-08-04 10:01:52

And then stops the service (2021-08-04 10:01:57), and restarts it (2021-08-04 10:02:03).

This resulted in the following error, which has NOT been readily reproducible:

ntdll.dll!RtlpUnWaitCriticalSection()
ntdll.dll!RtlEnterCriticalSection()
ntdll.dll!RtlEnterCriticalSection()
mysqld.exe!mysql_manager_submit()[sql_manager.cc:51]
mysqld.exe!rpl_slave_state::update()[rpl_gtid.cc:358]
mysqld.exe!rpl_load_gtid_slave_state()[rpl_rli.cc:1930]
mysqld.exe!binlog_background_thread()[log.cc:10026]
mysqld.exe!pthread_start()[my_winthread.c:62]
ucrtbase.dll!o_realloc_base()
KERNEL32.DLL!BaseThreadInitThunk()
ntdll.dll!RtlUserThreadStart()