| 1 |
# |
| 2 |
# WL#5569/7165 MTS |
| 3 |
# |
| 4 |
# The test verifies MTS recovery in cases SQL applier threads (workers) |
| 5 |
# have not completed all of their assignments at time they've been stopped. |
| 6 |
# The state of the slave server might become insonsistent with that of the master |
| 7 |
# 'cos there could be a number of event groups (transactions) uncommitted (gaps) followed |
| 8 |
# with that of committed. |
| 9 |
# The test proves that once the stop condition (e.g slave local error) is removed |
| 10 |
# replication restarts, the gaps are filled up, and the eventual slave server state |
| 11 |
# is consistent. |
| 12 |
# The test is MTS Logical_clock specific due to a mechanism of failure injection. |
| 13 |
|
| 14 |
# --let $rpl_topology= 1->2 |
| 15 |
# --let $rpl_debug= 1 |
| 16 |
# --source include/rpl_init.inc |
| 17 |
# --source include/rpl_default_connections.inc |
| 18 |
|
| 19 |
# Row format is chosen to ease the logged transaction size control |
| 20 |
--source include/have_binlog_format_row.inc |
| 21 |
--source include/force_myisam_default.inc |
| 22 |
--source include/have_myisam.inc |
| 23 |
--source include/not_group_replication_plugin.inc |
| 24 |
--source include/have_replica_parallel_type_logical_clock.inc |
| 25 |
--source include/master-slave.inc |
| 26 |
|
| 27 |
--source extra/rpl_tests/rpl_parallel_recovery_core.test |
| 28 |
|
| 29 |
--source include/rpl_end.inc |