mariadb/mysql-test/suite/galera/r/galera_ftwrl_concurrent.result
Teemu Ollakka ef7fc586ae MDEV-32282: Galera node remains paused after interleaving FTWRLs
After two concurrent FTWRL/UNLOCK TABLES, the node stays in paused state
and the following CREATE TABLE fails with

  ER_UNKNOWN_COM_ERROR (1047): Aborting TOI: Replication paused on
  node for FTWRL/BACKUP STAGE.

The cause is the use of global `wsrep_locked_seqno` to determine
if the node should be resumed on UNLOCK TABLES. In some executions
the `wsrep_locked_seqno` is cleared by the first UNLOCK TABLES
after the second FTWRL gets past `make_global_read_lock_block_commit()`.

As a fix, use `thd->wsrep_desynced_backup_stage` to determine
if the thread should resume the node on UNLOCK TABLES.

Add MTR test galera.galera_ftwrl_concurrent to reproduce the
race. The test contains also cases for BACKUP STAGE which
uses similar mechanism for desyncing and pausing the node.

Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
2023-10-26 22:13:54 +02:00

122 lines
4.2 KiB
Text

connection node_2;
connection node_1;
connect node_1_ctrl, 127.0.0.1, root, , test, $NODE_MYPORT_1;
SET SESSION wsrep_sync_wait=0;
#
# Case 1: FTWRL
#
connection node_1;
SET SESSION wsrep_sync_wait=0;
FLUSH TABLES WITH READ LOCK;
SHOW STATUS LIKE 'wsrep_local_state_comment';
Variable_name Value
wsrep_local_state_comment Donor/Desynced
SET SESSION debug_sync = "wsrep_unlock_global_read_lock_after_resume_and_resync SIGNAL unlock_ready WAIT_FOR unlock_continue";
UNLOCK TABLES;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR unlock_ready";
connect node_1_a, 127.0.0.1, root, , test, $NODE_MYPORT_1;
connection node_1_a;
SET SESSION debug_sync = "wsrep_global_read_lock_block_commit_after_pause SIGNAL lock_ready WAIT_FOR lock_continue";
FLUSH TABLES WITH READ LOCK;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR lock_ready";
SET debug_sync = "now SIGNAL unlock_continue";
connection node_1;
connection node_1_ctrl;
SET debug_sync = "now SIGNAL lock_continue";
connection node_1_a;
UNLOCK TABLES;
CREATE TABLE t1 (f1 INT PRIMARY KEY) ENGINE=InnoDB;
DROP TABLE t1;
connection node_1_ctrl;
SET debug_sync = "RESET";
#
# Case 2: BACKUP STAGE
#
connection node_1;
SET SESSION wsrep_sync_wait=0;
BACKUP STAGE START;
BACKUP STAGE BLOCK_DDL;
SHOW STATUS LIKE 'wsrep_local_state_comment';
Variable_name Value
wsrep_local_state_comment Donor/Desynced
SET SESSION debug_sync = "wsrep_backup_stage_after_resume_and_resync SIGNAL resume_and_resync_ready WAIT_FOR resume_and_resync_continue";
BACKUP STAGE END;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR resume_and_resync_ready";
connection node_1_a;
BACKUP STAGE START;
SET SESSION debug_sync = "wsrep_backup_stage_after_desync_and_pause SIGNAL desync_and_pause_ready WAIT_FOR desync_and_pause_continue";
BACKUP STAGE BLOCK_DDL;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR desync_and_pause_ready";
SET debug_sync = "now SIGNAL resume_and_resync_continue";
connection node_1;
connection node_1_ctrl;
SET debug_sync = "now SIGNAL desync_and_pause_continue";
connection node_1_a;
BACKUP STAGE END;
CREATE TABLE t1 (f1 INT PRIMARY KEY) ENGINE=InnoDB;
DROP TABLE t1;
connection node_1_ctrl;
SET debug_sync = "RESET";
#
# Case 3: FTWRL first, BACKUP STAGE second
#
connection node_1;
SET SESSION wsrep_sync_wait=0;
SET SESSION wsrep_sync_wait=0;
FLUSH TABLES WITH READ LOCK;
SHOW STATUS LIKE 'wsrep_local_state_comment';
Variable_name Value
wsrep_local_state_comment Donor/Desynced
SET SESSION debug_sync = "wsrep_unlock_global_read_lock_after_resume_and_resync SIGNAL unlock_ready WAIT_FOR unlock_continue";
UNLOCK TABLES;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR unlock_ready";
connection node_1_a;
BACKUP STAGE START;
SET SESSION debug_sync = "wsrep_backup_stage_after_desync_and_pause SIGNAL desync_and_pause_ready WAIT_FOR desync_and_pause_continue";
BACKUP STAGE BLOCK_DDL;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR desync_and_pause_ready";
SET debug_sync = "now SIGNAL unlock_continue";
connection node_1;
connection node_1_ctrl;
SET debug_sync = "now SIGNAL desync_and_pause_continue";
connection node_1_a;
BACKUP STAGE END;
CREATE TABLE t1 (f1 INT PRIMARY KEY) ENGINE=InnoDB;
DROP TABLE t1;
connection node_1_ctrl;
SET debug_sync = "RESET";
#
# Case 4: BACKUP STAGE first, then FTWRL
#
connection node_1;
SET SESSION wsrep_sync_wait=0;
BACKUP STAGE START;
BACKUP STAGE BLOCK_DDL;
SHOW STATUS LIKE 'wsrep_local_state_comment';
Variable_name Value
wsrep_local_state_comment Donor/Desynced
SET SESSION debug_sync = "wsrep_backup_stage_after_resume_and_resync SIGNAL resume_and_resync_ready WAIT_FOR resume_and_resync_continue";
BACKUP STAGE END;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR resume_and_resync_ready";
connection node_1_a;
SET SESSION debug_sync = "wsrep_global_read_lock_block_commit_after_pause SIGNAL lock_ready WAIT_FOR lock_continue";
FLUSH TABLES WITH READ LOCK;
connection node_1_ctrl;
SET debug_sync = "now WAIT_FOR lock_ready";
SET debug_sync = "now SIGNAL resume_and_resync_continue";
connection node_1;
connection node_1_ctrl;
SET debug_sync = "now SIGNAL lock_continue";
connection node_1_a;
UNLOCK TABLES;
CREATE TABLE t1 (f1 INT PRIMARY KEY) ENGINE=InnoDB;
DROP TABLE t1;
connection node_1_ctrl;
SET debug_sync = "RESET";