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!
* some of these tests run just fine with InnoDB:
-> s/have_xtradb/have_innodb/
* sys_var tests did basic tests for xtradb only variables
-> remove them, they're useless anyway (sysvar_innodb does it better)
* multi_update had innodb specific tests
-> move to multi_update_innodb.test
Relax the number-of-XA-engines check on recovery. Allow *more*
engines to be present than absolutely necessary, extra engines
cannot affect ACID guarantees of the recovery process.
As a bonus, 10.0->crash->10.1 upgrade won't complain about
wsrep being a new XA storge engine.
* handler::get_auto_increment() was not expecting any errors from the storage engine.
That was wrong, errors could happen.
* ha_partition::get_auto_increment() was doing index lookups in partition under a mutex.
This was redundant (engine transaction isolation was covering that anyway)
and harmful (causing deadlocks).
The problem was caused by the following scenario:
- range optimizer picks an index IDX1 which doesn't match the ORDER BY ...
LIMIT clause.
- test_if_skip_sort_order() decides to switch to index IDX2 which matches
the ORDER BY ... LIMIT.
- it runs SQL_SELECT::test_quick_select() for the second time to produce
an quick select for IDX2.
- However, test_quick_select() would figure that full index scan on IDX1
is still cheaper (its calculations ignore the LIMIT n).
Fixed this by
- passing force_quick_range=true to test_quick_select()
- in test_quick_select, don't consider full index scans if the mentioned
parameter is true.
Numerous changes in .result files are caused by test_quick_select() being
run after "early/late NULLs filtering" feature has injected NOT NULL
condition.
- Key_value_records_iterator::get_next() should pass pointer to the key
to handler->ha_index_next_same(). Because of a typo bug, pointer-to-pointer
was passed instead in certain cases.
mysql-test/r/create_or_replace.result:
More tests for create or replace
mysql-test/t/create_or_replace.test:
More tests for create or replace
sql/log.cc:
Don't use binlog_hton if binlog is not enabmed
sql/sql_base.cc:
We have to call restart_trans_for_tables also if tables where not locked with LOCK TABLES.
If not, we will get a crash in TokuDB
sql/sql_insert.cc:
Don't call binlog_reset_cache() if we don't have binary log open
sql/sql_table.cc:
Don't log to binary log if not open
Better test if we where using create or replace ... select
storage/tokudb/mysql-test/tokudb_mariadb/r/create_or_replace.result:
More tests for create or replace
storage/tokudb/mysql-test/tokudb_mariadb/t/create_or_replace.test:
More tests for create or replace
MDEV-5839 TokuDB tables not properly cleaned on DROP DATABASE
TokuDB does not support discover_table_names() and writes no files
in the database directory, so automatic filename-based
discover_table_names() doesn't work either. So, it must force .frm
file to disk in ::create()
Made stopping of slave more robust
Fixed tokudb test cases that gave different results between runs
Speed up some slow tokudb tests by adding begin ... commit
mysql-test/extra/rpl_tests/rpl_stop_slave.test:
Ensure that slaves are properly synced before they are stopped.
(Otherwise some tests results will be different between runs)
storage/innobase/buf/buf0buf.cc:
Fixed compiler warning
storage/tokudb/mysql-test/tokudb/r/cluster_filter_unpack_varchar_and_int_hidden.result:
Test case could be solved with index or range scan.
storage/tokudb/mysql-test/tokudb/t/cluster_filter_unpack_varchar_and_int_hidden.test:
Test case could be solved with index or range scan.
storage/tokudb/mysql-test/tokudb_bugs/r/5733_innodb.result:
Speed up test by adding begin...commit
storage/tokudb/mysql-test/tokudb_bugs/r/5733_tokudb.result:
Speed up test by adding begin...commit
storage/tokudb/mysql-test/tokudb_bugs/t/5733_innodb.test:
Speed up test by adding begin...commit
storage/tokudb/mysql-test/tokudb_bugs/t/5733_tokudb.test:
Speed up test by adding begin...commit
storage/tokudb/mysql-test/tokudb_mariadb/r/compression.result:
Added drop table (safety)
storage/tokudb/mysql-test/tokudb_mariadb/t/compression.test:
Added drop table (safety)
- the problem was caused by EXPLAIN INSERT SELECT. For that statement,
the code would call select_insert::prepare2(), which would call
handler->ha_start_bulk_insert(). The corresponding handler->end_bulk_insert()
call is made from select_insert::send_eof or select_insert::abort_result_set
which are never called for EXPLAIN INSERT SELECT.
- Fixed by re-using approach of mysql-5.6: don't call ha_start_bulk_insert() if
we are in EXPLAIN.
