mirror of
https://github.com/MariaDB/server.git
synced 2025-11-28 17:36:30 +03:00
c8ae357341
Lifted long standing limitation to the XA of rolling it back at the
transaction's
connection close even if the XA is prepared.
Prepared XA-transaction is made to sustain connection close or server
restart.
The patch consists of
- binary logging extension to write prepared XA part of
transaction signified with
its XID in a new XA_prepare_log_event. The concusion part -
with Commit or Rollback decision - is logged separately as
Query_log_event.
That is in the binlog the XA consists of two separate group of
events.
That makes the whole XA possibly interweaving in binlog with
other XA:s or regular transaction but with no harm to
replication and data consistency.
Gtid_log_event receives two more flags to identify which of the
two XA phases of the transaction it represents. With either flag
set also XID info is added to the event.
When binlog is ON on the server XID::formatID is
constrained to 4 bytes.
- engines are made aware of the server policy to keep up user
prepared XA:s so they (Innodb, rocksdb) don't roll them back
anymore at their disconnect methods.
- slave applier is refined to cope with two phase logged XA:s
including parallel modes of execution.
This patch does not address crash-safe logging of the new events which
is being addressed by MDEV-21469.
CORNER CASES: read-only, pure myisam, binlog-*, @@skip_log_bin, etc
Are addressed along the following policies.
1. The read-only at reconnect marks XID to fail for future
completion with ER_XA_RBROLLBACK.
2. binlog-* filtered XA when it changes engine data is regarded as
loggable even when nothing got cached for binlog. An empty
XA-prepare group is recorded. Consequent Commit-or-Rollback
succeeds in the Engine(s) as well as recorded into binlog.
3. The same applies to the non-transactional engine XA.
4. @@skip_log_bin=OFF does not record anything at XA-prepare
(obviously), but the completion event is recorded into binlog to
admit inconsistency with slave.
The following actions are taken by the patch.
At XA-prepare:
when empty binlog cache - don't do anything to binlog if RO,
otherwise write empty XA_prepare (assert(binlog-filter case)).
At Disconnect:
when Prepared && RO (=> no binlogging was done)
set Xid_cache_element::error := ER_XA_RBROLLBACK
*keep* XID in the cache, and rollback the transaction.
At XA-"complete":
Discover the error, if any don't binlog the "complete",
return the error to the user.
Kudos
-----
Alexey Botchkov took to drive this work initially.
Sergei Golubchik, Sergei Petrunja, Marko Mäkelä provided a number of
good recommendations.
Sergei Voitovich made a magnificent review and improvements to the code.
They all deserve a bunch of thanks for making this work done!
117 lines
3.8 KiB
Plaintext
117 lines
3.8 KiB
Plaintext
source include/have_binlog_format_row.inc;
|
|
source include/have_innodb.inc;
|
|
source include/master-slave.inc;
|
|
|
|
call mtr.add_suppression("Deadlock found");
|
|
call mtr.add_suppression("Can't find record in 't.'");
|
|
|
|
connection master;
|
|
CREATE TABLE t1 (a INT PRIMARY KEY, b INT) ENGINE=innodb;
|
|
INSERT INTO t1 VALUES (1,1), (2,2), (3,3), (4,4);
|
|
sync_slave_with_master;
|
|
SHOW STATUS LIKE 'Slave_retried_transactions';
|
|
# since bug#31552/31609 idempotency is not default any longer. In order
|
|
# the following UPDATE t1 to pass the mode is switched temprorarily
|
|
set @@global.slave_exec_mode= 'IDEMPOTENT';
|
|
UPDATE t1 SET a = 5, b = 47 WHERE a = 1;
|
|
SELECT * FROM t1 ORDER BY a;
|
|
connection master;
|
|
UPDATE t1 SET a = 5, b = 5 WHERE a = 1;
|
|
SELECT * FROM t1 ORDER BY a;
|
|
#SHOW BINLOG EVENTS;
|
|
sync_slave_with_master;
|
|
set @@global.slave_exec_mode= default;
|
|
SHOW STATUS LIKE 'Slave_retried_transactions';
|
|
SELECT * FROM t1 ORDER BY a;
|
|
source include/check_slave_is_running.inc;
|
|
|
|
connection slave;
|
|
call mtr.add_suppression("Slave SQL.*Could not execute Update_rows event on table test.t1");
|
|
call mtr.add_suppression("Slave SQL for channel '': worker thread retried transaction");
|
|
call mtr.add_suppression("The slave coordinator and worker threads are stopped");
|
|
#
|
|
# Bug#24764800 REPLICATION FAILING ON SLAVE WITH XAER_RMFAIL ERROR
|
|
#
|
|
# Verify that a temporary failing replicated xa transaction completes
|
|
# upon slave applier restart after previous
|
|
# @@global.slave_transaction_retries number of retries in vain.
|
|
#
|
|
connection slave;
|
|
|
|
set @save_innodb_lock_wait_timeout=@@global.innodb_lock_wait_timeout;
|
|
set @save_slave_transaction_retries=@@global.slave_transaction_retries;
|
|
|
|
# Slave applier parameters for the failed retry
|
|
set @@global.innodb_lock_wait_timeout=1;
|
|
set @@global.slave_transaction_retries=2;
|
|
--source include/restart_slave_sql.inc
|
|
|
|
# Temporary error implement: a record is blocked by slave local trx
|
|
connection slave1;
|
|
BEGIN;
|
|
INSERT INTO t1 SET a = 6, b = 7;
|
|
|
|
connection master;
|
|
INSERT INTO t1 SET a = 99, b = 99; # slave applier warm up trx
|
|
XA START 'xa1';
|
|
INSERT INTO t1 SET a = 6, b = 6; # this record eventually must be found on slave
|
|
XA END 'xa1';
|
|
XA PREPARE 'xa1';
|
|
|
|
connection slave;
|
|
# convert_error(ER_LOCK_WAIT_TIMEOUT)
|
|
--let $err_timeout= 1205
|
|
# convert_error(ER_LOCK_DEADLOCK)
|
|
--let $err_deadlock= 1213
|
|
--let $slave_sql_errno=$err_deadlock,$err_timeout
|
|
--let $show_slave_sql_error=
|
|
--source include/wait_for_slave_sql_error.inc
|
|
|
|
# b. Slave applier parameters for successful retry after restart
|
|
set @@global.innodb_lock_wait_timeout=1;
|
|
set @@global.slave_transaction_retries=100;
|
|
|
|
--source include/restart_slave_sql.inc
|
|
|
|
--let $last_retries= query_get_value(SHOW GLOBAL STATUS LIKE 'Slave_retried_transactions', Value, 1)
|
|
--let $status_type=GLOBAL
|
|
--let $status_var=Slave_retried_transactions
|
|
--let $status_var_value=`SELECT 1 + $last_retries`
|
|
--let $$status_var_comparsion= >
|
|
--source include/wait_for_status_var.inc
|
|
|
|
# Release the record after just one retry
|
|
connection slave1;
|
|
ROLLBACK;
|
|
|
|
connection master;
|
|
XA COMMIT 'xa1';
|
|
|
|
--source include/sync_slave_sql_with_master.inc
|
|
|
|
# Proof of correctness: the committed XA is on the slave
|
|
connection slave;
|
|
--let $assert_text=XA transaction record must be in the table
|
|
--let $assert_cond=count(*)=1 FROM t1 WHERE a=6 AND b=6
|
|
--source include/assert.inc
|
|
|
|
# Bug#24764800 cleanup:
|
|
set @@global.innodb_lock_wait_timeout=@save_innodb_lock_wait_timeout;
|
|
set @@global.slave_transaction_retries= @save_slave_transaction_retries;
|
|
|
|
#
|
|
# Total cleanup:
|
|
#
|
|
connection master;
|
|
DROP TABLE t1;
|
|
--sync_slave_with_master
|
|
--connection master
|
|
|
|
# We must wait for the slave to stop.
|
|
# Otherwise the warnings in the error log about deadlock may be written to
|
|
# the error log only during shutdown, and currently the suppression of
|
|
# "Deadlock found" set in this test case is not effective during server
|
|
# shutdown.
|
|
|
|
--source include/rpl_end.inc
|