mariadb/sql/event_scheduler.h

156 lines
3.1 KiB
C
Raw Normal View History

fix for bug #17619 Scheduler race conditions - Scheduler is either initialized at server start or never. Starting & stopping is now suspending & resuming. - The scheduler has clear OO interface - Now all calls to the scheduler are synchronous - GLOBAL event_scheduler uses thd::sys_var_tmp (see set_var.cc) - External API is encapsulated into class Events - Includes fixes for all comments of Kostja's review of 19.05.2005 Starting to merge into 5.1-release (5.1.10) and push BitKeeper/etc/ignore: Added libmysqld/event_scheduler.cc to the ignore list libmysqld/Makefile.am: executor -> scheduler mysql-test/r/events.result: update result mysql-test/r/events_bugs.result: update result mysql-test/r/events_logs_tests.result: update result mysql-test/r/events_microsec.result: update result mysql-test/r/events_scheduling.result: update result mysql-test/r/events_stress.result: update result mysql-test/t/disabled.def: enable these tests mysql-test/t/events.test: optimize the test a bit for speed, save some seconds runtime remove FULL from SHOW EVENTS mostly use I_S.EVENTS mysql-test/t/events_bugs.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_logs_tests.test: optimize the test a bit for speed mysql-test/t/events_microsec.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_scheduling.test: broader test mysql-test/t/events_stress.test: Rework the test to the new architecture of suspending/resuming. Use less events, no need for thousands, hundreds is still ok. sql/Makefile.am: executor -> scheduler sql/cmakelists.txt: executor -> scheduler sql/event.cc: - remove todo comments - remove unneded evex_queue abstraction functions - move events_init() and events_shutdown() from event_executor.cc to here - export db_create_event - remove evex_load_and_compile_event, part of class Event_scheduler - integrate the public interface found in event.h and used by sql_parse.cc to use the new class Event_scheduler. sql/event.h: - add COND_finished so if one thread kills a running event it waits on this - export callback event_timed_definer_equal, event_timed_identifier_equal(), event_timed_name_equal and event_timed_db_equal() to be used by Event_scheduler::drop_matching_events() - cleanup event.h - encapsulated all external interface into class Events sql/event_executor.cc: make it empty, will delete after that sql/event_priv.h: - more things in the private header - remove event queue abstraction functions. tightly bind to QUEUE - export privately db_drop_event, db_find_event, db_create_event() - made change_security_context() and restore_security_context() free functions sql/event_timed.cc: - fix calculation of time when ENDS is set (STARTS is always set) - during Event_timed::compile() set the right Security_ctx. Prevents a crash during Event_scheduler::load_events_from_db() - add Event_timed::kill_thread() - implement event_timed_*_equal() - made change_security_context() and restore_security_context() free functions. - Comments cleanups sql/lex.h: new word scheduler for SHOW SCHEDULER STATUS (available only debug builds) sql/log.cc: move these from event_scheduler.cc sql/mysql_priv.h: refactor kill_one_thread export sql_print_message_func and sql_print_message_handlers sql/mysqld.cc: In close_connections, called by kill_server() skip the main scheduler thread and use events_shutdown() for shutting down the scheduler, in the same manner it's done for RPL. Add a new value to --event-scheduler : 0 <- No scheduler available 1 <- Start with scheduler enabled 2 <- Start with scheduler suspended sql/repl_failsafe.cc: refactor thd::system_thread to be an enum sql/set_var.cc: move sys_var_event_executor::update() to set_var.cc executor -> scheduler use thd::sys_var_tmp sql/set_var.h: executor -> scheduler sql/share/errmsg.txt: 3 new error messages sql/sql_class.cc: refactor thd::system_thread to be an enum . more type-safety sql/sql_class.h: refactor thd::system_thread to be an enum . more type-safety sql/sql_db.cc: get the error from evex_drop_schema_events sql/sql_error.h: export warning_level_names sql/sql_lex.h: new command SHOW SCHEDULER STATUS, available only in debug build and for debug purposes. sql/sql_parse.cc: refactor kill_one_thread() -> does the *dirty* work, and sql_kill just the reporting. add handler for SQLCOM_SHOW_SCHEDULER_STATUS sql/sql_show.cc: fix verbosity handling (this will be obsoleted anyway by the fix for 17394). sql/sql_yacc.yy: remove FULL from SHOW EVENTS add SHOW SCHEDULER STATUS in debug builds sql/table.cc: Fix valgrind warning.
2006-05-22 20:46:13 +02:00
#ifndef _EVENT_SCHEDULER_H_
#define _EVENT_SCHEDULER_H_
/* Copyright (C) 2004-2006 MySQL AB
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
Many files: Changed header to GPL version 2 only client/mysqlslap.c: Changed header to GPL version 2 only include/atomic/nolock.h: Changed header to GPL version 2 only include/atomic/rwlock.h: Changed header to GPL version 2 only include/atomic/x86-gcc.h: Changed header to GPL version 2 only include/atomic/x86-msvc.h: Changed header to GPL version 2 only include/my_atomic.h: Changed header to GPL version 2 only include/my_trie.h: Changed header to GPL version 2 only include/my_vle.h: Changed header to GPL version 2 only include/mysql/plugin.h: Changed header to GPL version 2 only mysys/my_atomic.c: Changed header to GPL version 2 only mysys/my_getncpus.c: Changed header to GPL version 2 only mysys/my_memmem.c: Changed header to GPL version 2 only mysys/my_vle.c: Changed header to GPL version 2 only mysys/trie.c: Changed header to GPL version 2 only plugin/Makefile.am: Changed header to GPL version 2 only server-tools/instance-manager/IMService.h: Changed header to GPL version 2 only server-tools/instance-manager/WindowsService.h: Changed header to GPL version 2 only server-tools/instance-manager/exit_codes.h: Changed header to GPL version 2 only server-tools/instance-manager/user_management_commands.h: Changed header to GPL version 2 only sql/authors.h: Changed header to GPL version 2 only sql/contributors.h: Changed header to GPL version 2 only sql/event_data_objects.cc: Changed header to GPL version 2 only sql/event_data_objects.h: Changed header to GPL version 2 only sql/event_db_repository.cc: Changed header to GPL version 2 only sql/event_db_repository.h: Changed header to GPL version 2 only sql/event_queue.cc: Changed header to GPL version 2 only sql/event_queue.h: Changed header to GPL version 2 only sql/event_scheduler.cc: Changed header to GPL version 2 only sql/event_scheduler.h: Changed header to GPL version 2 only sql/events.cc: Changed header to GPL version 2 only sql/events.h: Changed header to GPL version 2 only sql/ha_ndbcluster_binlog.cc: Changed header to GPL version 2 only sql/ha_ndbcluster_binlog.h: Changed header to GPL version 2 only sql/ha_ndbcluster_tables.h: Changed header to GPL version 2 only sql/ha_partition.cc: Changed header to GPL version 2 only sql/ha_partition.h: Changed header to GPL version 2 only sql/item_xmlfunc.cc: Changed header to GPL version 2 only sql/item_xmlfunc.h: Changed header to GPL version 2 only sql/log.h: Changed header to GPL version 2 only sql/partition_element.h: Changed header to GPL version 2 only sql/partition_info.cc: Changed header to GPL version 2 only sql/partition_info.h: Changed header to GPL version 2 only sql/rpl_filter.cc: Changed header to GPL version 2 only sql/rpl_filter.h: Changed header to GPL version 2 only sql/rpl_injector.cc: Changed header to GPL version 2 only sql/rpl_injector.h: Changed header to GPL version 2 only sql/rpl_mi.cc: Changed header to GPL version 2 only sql/rpl_mi.h: Changed header to GPL version 2 only sql/rpl_rli.cc: Changed header to GPL version 2 only sql/rpl_rli.h: Changed header to GPL version 2 only sql/rpl_tblmap.cc: Changed header to GPL version 2 only sql/rpl_tblmap.h: Changed header to GPL version 2 only sql/rpl_utility.cc: Changed header to GPL version 2 only sql/rpl_utility.h: Changed header to GPL version 2 only sql/sql_binlog.cc: Changed header to GPL version 2 only sql/sql_partition.cc: Changed header to GPL version 2 only sql/sql_partition.h: Changed header to GPL version 2 only sql/sql_plugin.cc: Changed header to GPL version 2 only sql/sql_plugin.h: Changed header to GPL version 2 only sql/sql_servers.cc: Changed header to GPL version 2 only sql/sql_servers.h: Changed header to GPL version 2 only sql/sql_tablespace.cc: Changed header to GPL version 2 only sql/sql_yacc.yy.bak: Changed header to GPL version 2 only storage/Makefile.am: Changed header to GPL version 2 only storage/archive/Makefile.am: Changed header to GPL version 2 only storage/blackhole/Makefile.am: Changed header to GPL version 2 only storage/csv/Makefile.am: Changed header to GPL version 2 only storage/example/Makefile.am: Changed header to GPL version 2 only storage/federated/Makefile.am: Changed header to GPL version 2 only storage/innobase/handler/Makefile.am: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/AllocNodeId.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/CreateFilegroup.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/CreateFilegroupImpl.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/CreateObj.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/DictObjOp.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/DihFragCount.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/DropFilegroup.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/DropFilegroupImpl.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/DropObj.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/Extent.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/LgmanContinueB.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/PgmanContinueB.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/RestoreContinueB.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/RestoreImpl.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/RouteOrd.hpp: Changed header to GPL version 2 only storage/ndb/include/kernel/signaldata/TsmanContinueB.hpp: Changed header to GPL version 2 only storage/ndb/include/ndbapi/NdbIndexStat.hpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/mgmapi_logevent/mgmapi_logevent.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/mgmapi_logevent_dual/mgmapi_logevent_dual.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_async/ndbapi_async.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_async1/ndbapi_async1.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_event/ndbapi_event.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_retries/ndbapi_retries.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_scan/ndbapi_scan.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_simple/ndbapi_simple.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_simple_dual/ndbapi_simple_dual.cpp: Changed header to GPL version 2 only storage/ndb/ndbapi-examples/ndbapi_simple_index/ndbapi_simple_index.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbdih/printSysfile.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/DbtupDiskAlloc.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/DbtupVarAlloc.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/Undo_buffer.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/Undo_buffer.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/tuppage.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtup/tuppage.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/dbtux/DbtuxStat.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/diskpage.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/lgman.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/lgman.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/pgman.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/pgman.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/print_file.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/record_types.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/restore.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/restore.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/tsman.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/blocks/tsman.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/DLCFifoList.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/DLCHashTable.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/DynArr256.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/DynArr256.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/KeyTable2Ref.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/LinearPool.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/NdbdSuperPool.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/NdbdSuperPool.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/Pool.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/Pool.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/RWPool.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/RWPool.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/Rope.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/SLFifoList.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/WOPool.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/WOPool.hpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/bench_pool.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/ndbd_malloc_impl.cpp: Changed header to GPL version 2 only storage/ndb/src/kernel/vm/ndbd_malloc_impl.hpp: Changed header to GPL version 2 only storage/ndb/src/mgmsrv/ParamInfo.cpp: Changed header to GPL version 2 only storage/ndb/src/ndbapi/NdbIndexStat.cpp: Changed header to GPL version 2 only storage/ndb/test/ndbapi/testIndexStat.cpp: Changed header to GPL version 2 only storage/ndb/test/tools/listen.cpp: Changed header to GPL version 2 only storage/ndb/tools/restore/ndb_nodegroup_map.h: Changed header to GPL version 2 only strings/my_strchr.c: Changed header to GPL version 2 only unittest/mysys/base64-t.c: Changed header to GPL version 2 only unittest/mysys/bitmap-t.c: Changed header to GPL version 2 only unittest/mysys/my_atomic-t.c: Changed header to GPL version 2 only unittest/mytap/tap.c: Changed header to GPL version 2 only unittest/mytap/tap.h: Changed header to GPL version 2 only win/Makefile.am: Changed header to GPL version 2 only
2006-12-27 02:23:51 +01:00
the Free Software Foundation; version 2 of the License.
fix for bug #17619 Scheduler race conditions - Scheduler is either initialized at server start or never. Starting & stopping is now suspending & resuming. - The scheduler has clear OO interface - Now all calls to the scheduler are synchronous - GLOBAL event_scheduler uses thd::sys_var_tmp (see set_var.cc) - External API is encapsulated into class Events - Includes fixes for all comments of Kostja's review of 19.05.2005 Starting to merge into 5.1-release (5.1.10) and push BitKeeper/etc/ignore: Added libmysqld/event_scheduler.cc to the ignore list libmysqld/Makefile.am: executor -> scheduler mysql-test/r/events.result: update result mysql-test/r/events_bugs.result: update result mysql-test/r/events_logs_tests.result: update result mysql-test/r/events_microsec.result: update result mysql-test/r/events_scheduling.result: update result mysql-test/r/events_stress.result: update result mysql-test/t/disabled.def: enable these tests mysql-test/t/events.test: optimize the test a bit for speed, save some seconds runtime remove FULL from SHOW EVENTS mostly use I_S.EVENTS mysql-test/t/events_bugs.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_logs_tests.test: optimize the test a bit for speed mysql-test/t/events_microsec.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_scheduling.test: broader test mysql-test/t/events_stress.test: Rework the test to the new architecture of suspending/resuming. Use less events, no need for thousands, hundreds is still ok. sql/Makefile.am: executor -> scheduler sql/cmakelists.txt: executor -> scheduler sql/event.cc: - remove todo comments - remove unneded evex_queue abstraction functions - move events_init() and events_shutdown() from event_executor.cc to here - export db_create_event - remove evex_load_and_compile_event, part of class Event_scheduler - integrate the public interface found in event.h and used by sql_parse.cc to use the new class Event_scheduler. sql/event.h: - add COND_finished so if one thread kills a running event it waits on this - export callback event_timed_definer_equal, event_timed_identifier_equal(), event_timed_name_equal and event_timed_db_equal() to be used by Event_scheduler::drop_matching_events() - cleanup event.h - encapsulated all external interface into class Events sql/event_executor.cc: make it empty, will delete after that sql/event_priv.h: - more things in the private header - remove event queue abstraction functions. tightly bind to QUEUE - export privately db_drop_event, db_find_event, db_create_event() - made change_security_context() and restore_security_context() free functions sql/event_timed.cc: - fix calculation of time when ENDS is set (STARTS is always set) - during Event_timed::compile() set the right Security_ctx. Prevents a crash during Event_scheduler::load_events_from_db() - add Event_timed::kill_thread() - implement event_timed_*_equal() - made change_security_context() and restore_security_context() free functions. - Comments cleanups sql/lex.h: new word scheduler for SHOW SCHEDULER STATUS (available only debug builds) sql/log.cc: move these from event_scheduler.cc sql/mysql_priv.h: refactor kill_one_thread export sql_print_message_func and sql_print_message_handlers sql/mysqld.cc: In close_connections, called by kill_server() skip the main scheduler thread and use events_shutdown() for shutting down the scheduler, in the same manner it's done for RPL. Add a new value to --event-scheduler : 0 <- No scheduler available 1 <- Start with scheduler enabled 2 <- Start with scheduler suspended sql/repl_failsafe.cc: refactor thd::system_thread to be an enum sql/set_var.cc: move sys_var_event_executor::update() to set_var.cc executor -> scheduler use thd::sys_var_tmp sql/set_var.h: executor -> scheduler sql/share/errmsg.txt: 3 new error messages sql/sql_class.cc: refactor thd::system_thread to be an enum . more type-safety sql/sql_class.h: refactor thd::system_thread to be an enum . more type-safety sql/sql_db.cc: get the error from evex_drop_schema_events sql/sql_error.h: export warning_level_names sql/sql_lex.h: new command SHOW SCHEDULER STATUS, available only in debug build and for debug purposes. sql/sql_parse.cc: refactor kill_one_thread() -> does the *dirty* work, and sql_kill just the reporting. add handler for SQLCOM_SHOW_SCHEDULER_STATUS sql/sql_show.cc: fix verbosity handling (this will be obsoleted anyway by the fix for 17394). sql/sql_yacc.yy: remove FULL from SHOW EVENTS add SHOW SCHEDULER STATUS in debug builds sql/table.cc: Fix valgrind warning.
2006-05-22 20:46:13 +02:00
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
/**
@addtogroup Event_Scheduler
@{
*/
A set of changes aiming to make the Event Scheduler more user-friendly when there are no up-to-date system tables to support it: - initialize the scheduler before reporting "Ready for connections". This ensures that warnings, if any, are printed before "Ready for connections", and this message is not mangled. - do not abort the scheduler if there are no system tables - check the tables once at start up, remember the status and disable the scheduler if the tables are not up to date. If one attempts to use the scheduler with bad tables, issue an error message. - clean up the behaviour of the module under LOCK TABLES and pre-locking mode - make sure implicit commit of Events DDL works as expected. - add more tests Collateral clean ups in the events code. This patch fixes Bug#23631 Events: SHOW VARIABLES doesn't work when mysql.event is damaged mysql-test/r/events.result: Update results. mysql-test/r/events_bugs.result: Update results. mysql-test/r/events_restart_phase1.result: Update results. mysql-test/r/events_restart_phase2.result: Update results. mysql-test/r/events_restart_phase3.result: Update results. mysql-test/r/events_scheduling.result: Update results. mysql-test/r/events_time_zone.result: Update results. mysql-test/t/events.test: Add new tests for tampering with mysql.event and some more tests for sub-statements, LOCK TABLES mode and pre-locking. mysql-test/t/events_bugs.test: Move the non-concurrent part of test for Bug 16420 to this file. mysql-test/t/events_restart_phase1.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase2.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase3.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_scheduling.test: Add more coverage for event_scheduler global variable. mysql-test/t/events_time_zone.test: Move the non-concurrent part of the tests for Bug 16420 to events_bugs.test sql/event_data_objects.cc: Move update_timing_fields functionality to Event_db_repository. Make loading of events from a table record more robust to tampering with the table - now we do not check mysql.event on every table open. sql/event_data_objects.h: Cleanup. sql/event_db_repository.cc: Now Event_db_repository is responsible for table I/O only. All the logic of events DDL is handled outside, in Events class please refer to the added test coverage to see how this change affected the behavior of Event Scheduler. Dependency on sp_head.h and sp.h removed. Make this module robust to tweaks with mysql.event table. Move check_system_tables from events.cc to this file sql/event_db_repository.h: Cleanup declarations (remove unused ones, change return type to bool from int). sql/event_queue.cc: Update to adapt to the new start up scheme of the Event Scheduler. sql/event_queue.h: Cleanup declarations. sql/event_scheduler.cc: Make all the error messages uniform: [SEVERITY] Event Scheduler: [user][schema.event] message Using append_identifier for error logging was an overkill - we may need it only if the system character set may have NUL (null character) as part of a valid identifier, this is currently never the case, whereas additional quoting did not look nice in the log. sql/event_scheduler.h: Cleanup the headers. sql/events.cc: Use a different start up procedure of Event Scheduler: - at start up, try to check the system tables first. If they are not up-to-date, disable the scheduler. - try to load all the active events. In case of a load error, abort start up. - do not parse an event on start up. Parsing only gives some information about event validity, but far not all. Consolidate the business logic of Events DDL in this module. Now opt_event_scheduler may change after start up and thus is protected by LOCK_event_metadata mutex. sql/events.h: Use all-static-data-members approach to implement Singleton pattern. sql/mysqld.cc: New invocation scheme of Events. Move some logic to events.cc. Initialize the scheduler before reporting "Ready for connections". sql/set_var.cc: Clean up sys_var_thd_sql_mode::symbolic_mode_representation to work with a LEX_STRING. Move more logic related to @@events_scheduler global variable to Events module. sql/set_var.h: Update declarations. sql/share/errmsg.txt: If someone tampered with mysql.event table after the server has started we no longer give him/her a complete report what was actually broken. Do not send the user to look at the error log in such case, as there is nothing there (check_table_intact is not executed). sql/sp_head.cc: Update to a new declaration of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_db.cc: New invocation scheme of Events module. sql/sql_parse.cc: Move more logic to Events module. Make sure that we are consistent in the way access rights are checked for Events DDL: always after committing the current transaction and checking the system tables. sql/sql_show.cc: Update to the new declarations of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_test.cc: New invocation scheme of events. sql/table.cc: mysql.event is a system table. Update check_table_intact to be concurrent, more verbose, and less smart. sql/table.h: Add a helper method. mysql-test/r/events_trans.result: New BitKeeper file ``mysql-test/r/events_trans.result'' mysql-test/t/events_trans.test: New BitKeeper file ``mysql-test/t/events_trans.test'': test cases for Event Scheduler that require a transactional storage engine.
2007-04-05 13:24:34 +02:00
/**
@file
Declarations of the scheduler thread class
and related functionality.
This file is internal to Event_Scheduler module. Please do not
include it directly. All public declarations of Event_Scheduler
module are in events.h and event_data_objects.h.
*/
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
class Event_queue;
class Event_job_data;
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
class Event_db_repository;
class Events;
WL #3337 (Events new architecture) Final stroke, events should be loaded from disk on server startup. Also check the validity of their bodies if possible during loading. sql/event_data_objects.cc: Remove Event_job_data::free_sp(), move the code to the destructor Change the way we change the security context Steal some code from sql_parse.cc sql/event_data_objects.h: Remove free_sp() Make compile() public, to be used when booting for verifying the integrity of mysql.event sql/event_queue.cc: Make the queue load events from disk on server boot. Compile and thus check for integrity the events. sql/event_queue.h: shift methods around. add queue_loaded boolean. sql/event_scheduler.cc: Rename init_event_thread() to pre_init_event_thread() and make it more generic. Add post_init_event_thread() Export these two as well as deinit_event_thread(). Now it is quite easy to write code to spawn a new event thread whenever needed. sql/event_scheduler.h: export pre_init_event_thread(), post_init_event_thread() and deinit_event_thread() to simplify writing of thread functions. sql/events.cc: Events::init() returns only one error code, then make it bool sql/events.h: Events::init() returns only one error code, then make it bool sql/mysqld.cc: Check the return code of Events::init() sql/sp_head.cc: Add trace info sql/sql_class.cc: Reorganize thd::change_security_context() to load main_security_ctx sql/sql_class.h: Reorganize thd::change_security_context() to load main_security_ctx sql/sql_lex.cc: Initialize lex->spname sql/sql_yacc.yy: Add a comment
2006-07-13 10:59:58 +02:00
void
pre_init_event_thread(THD* thd);
bool
post_init_event_thread(THD* thd);
void
deinit_event_thread(THD *thd);
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
class Event_worker_thread
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
{
public:
static void
A set of changes aiming to make the Event Scheduler more user-friendly when there are no up-to-date system tables to support it: - initialize the scheduler before reporting "Ready for connections". This ensures that warnings, if any, are printed before "Ready for connections", and this message is not mangled. - do not abort the scheduler if there are no system tables - check the tables once at start up, remember the status and disable the scheduler if the tables are not up to date. If one attempts to use the scheduler with bad tables, issue an error message. - clean up the behaviour of the module under LOCK TABLES and pre-locking mode - make sure implicit commit of Events DDL works as expected. - add more tests Collateral clean ups in the events code. This patch fixes Bug#23631 Events: SHOW VARIABLES doesn't work when mysql.event is damaged mysql-test/r/events.result: Update results. mysql-test/r/events_bugs.result: Update results. mysql-test/r/events_restart_phase1.result: Update results. mysql-test/r/events_restart_phase2.result: Update results. mysql-test/r/events_restart_phase3.result: Update results. mysql-test/r/events_scheduling.result: Update results. mysql-test/r/events_time_zone.result: Update results. mysql-test/t/events.test: Add new tests for tampering with mysql.event and some more tests for sub-statements, LOCK TABLES mode and pre-locking. mysql-test/t/events_bugs.test: Move the non-concurrent part of test for Bug 16420 to this file. mysql-test/t/events_restart_phase1.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase2.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase3.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_scheduling.test: Add more coverage for event_scheduler global variable. mysql-test/t/events_time_zone.test: Move the non-concurrent part of the tests for Bug 16420 to events_bugs.test sql/event_data_objects.cc: Move update_timing_fields functionality to Event_db_repository. Make loading of events from a table record more robust to tampering with the table - now we do not check mysql.event on every table open. sql/event_data_objects.h: Cleanup. sql/event_db_repository.cc: Now Event_db_repository is responsible for table I/O only. All the logic of events DDL is handled outside, in Events class please refer to the added test coverage to see how this change affected the behavior of Event Scheduler. Dependency on sp_head.h and sp.h removed. Make this module robust to tweaks with mysql.event table. Move check_system_tables from events.cc to this file sql/event_db_repository.h: Cleanup declarations (remove unused ones, change return type to bool from int). sql/event_queue.cc: Update to adapt to the new start up scheme of the Event Scheduler. sql/event_queue.h: Cleanup declarations. sql/event_scheduler.cc: Make all the error messages uniform: [SEVERITY] Event Scheduler: [user][schema.event] message Using append_identifier for error logging was an overkill - we may need it only if the system character set may have NUL (null character) as part of a valid identifier, this is currently never the case, whereas additional quoting did not look nice in the log. sql/event_scheduler.h: Cleanup the headers. sql/events.cc: Use a different start up procedure of Event Scheduler: - at start up, try to check the system tables first. If they are not up-to-date, disable the scheduler. - try to load all the active events. In case of a load error, abort start up. - do not parse an event on start up. Parsing only gives some information about event validity, but far not all. Consolidate the business logic of Events DDL in this module. Now opt_event_scheduler may change after start up and thus is protected by LOCK_event_metadata mutex. sql/events.h: Use all-static-data-members approach to implement Singleton pattern. sql/mysqld.cc: New invocation scheme of Events. Move some logic to events.cc. Initialize the scheduler before reporting "Ready for connections". sql/set_var.cc: Clean up sys_var_thd_sql_mode::symbolic_mode_representation to work with a LEX_STRING. Move more logic related to @@events_scheduler global variable to Events module. sql/set_var.h: Update declarations. sql/share/errmsg.txt: If someone tampered with mysql.event table after the server has started we no longer give him/her a complete report what was actually broken. Do not send the user to look at the error log in such case, as there is nothing there (check_table_intact is not executed). sql/sp_head.cc: Update to a new declaration of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_db.cc: New invocation scheme of Events module. sql/sql_parse.cc: Move more logic to Events module. Make sure that we are consistent in the way access rights are checked for Events DDL: always after committing the current transaction and checking the system tables. sql/sql_show.cc: Update to the new declarations of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_test.cc: New invocation scheme of events. sql/table.cc: mysql.event is a system table. Update check_table_intact to be concurrent, more verbose, and less smart. sql/table.h: Add a helper method. mysql-test/r/events_trans.result: New BitKeeper file ``mysql-test/r/events_trans.result'' mysql-test/t/events_trans.test: New BitKeeper file ``mysql-test/t/events_trans.test'': test cases for Event Scheduler that require a transactional storage engine.
2007-04-05 13:24:34 +02:00
init(Event_db_repository *db_repository_arg)
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
{
A set of changes aiming to make the Event Scheduler more user-friendly when there are no up-to-date system tables to support it: - initialize the scheduler before reporting "Ready for connections". This ensures that warnings, if any, are printed before "Ready for connections", and this message is not mangled. - do not abort the scheduler if there are no system tables - check the tables once at start up, remember the status and disable the scheduler if the tables are not up to date. If one attempts to use the scheduler with bad tables, issue an error message. - clean up the behaviour of the module under LOCK TABLES and pre-locking mode - make sure implicit commit of Events DDL works as expected. - add more tests Collateral clean ups in the events code. This patch fixes Bug#23631 Events: SHOW VARIABLES doesn't work when mysql.event is damaged mysql-test/r/events.result: Update results. mysql-test/r/events_bugs.result: Update results. mysql-test/r/events_restart_phase1.result: Update results. mysql-test/r/events_restart_phase2.result: Update results. mysql-test/r/events_restart_phase3.result: Update results. mysql-test/r/events_scheduling.result: Update results. mysql-test/r/events_time_zone.result: Update results. mysql-test/t/events.test: Add new tests for tampering with mysql.event and some more tests for sub-statements, LOCK TABLES mode and pre-locking. mysql-test/t/events_bugs.test: Move the non-concurrent part of test for Bug 16420 to this file. mysql-test/t/events_restart_phase1.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase2.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase3.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_scheduling.test: Add more coverage for event_scheduler global variable. mysql-test/t/events_time_zone.test: Move the non-concurrent part of the tests for Bug 16420 to events_bugs.test sql/event_data_objects.cc: Move update_timing_fields functionality to Event_db_repository. Make loading of events from a table record more robust to tampering with the table - now we do not check mysql.event on every table open. sql/event_data_objects.h: Cleanup. sql/event_db_repository.cc: Now Event_db_repository is responsible for table I/O only. All the logic of events DDL is handled outside, in Events class please refer to the added test coverage to see how this change affected the behavior of Event Scheduler. Dependency on sp_head.h and sp.h removed. Make this module robust to tweaks with mysql.event table. Move check_system_tables from events.cc to this file sql/event_db_repository.h: Cleanup declarations (remove unused ones, change return type to bool from int). sql/event_queue.cc: Update to adapt to the new start up scheme of the Event Scheduler. sql/event_queue.h: Cleanup declarations. sql/event_scheduler.cc: Make all the error messages uniform: [SEVERITY] Event Scheduler: [user][schema.event] message Using append_identifier for error logging was an overkill - we may need it only if the system character set may have NUL (null character) as part of a valid identifier, this is currently never the case, whereas additional quoting did not look nice in the log. sql/event_scheduler.h: Cleanup the headers. sql/events.cc: Use a different start up procedure of Event Scheduler: - at start up, try to check the system tables first. If they are not up-to-date, disable the scheduler. - try to load all the active events. In case of a load error, abort start up. - do not parse an event on start up. Parsing only gives some information about event validity, but far not all. Consolidate the business logic of Events DDL in this module. Now opt_event_scheduler may change after start up and thus is protected by LOCK_event_metadata mutex. sql/events.h: Use all-static-data-members approach to implement Singleton pattern. sql/mysqld.cc: New invocation scheme of Events. Move some logic to events.cc. Initialize the scheduler before reporting "Ready for connections". sql/set_var.cc: Clean up sys_var_thd_sql_mode::symbolic_mode_representation to work with a LEX_STRING. Move more logic related to @@events_scheduler global variable to Events module. sql/set_var.h: Update declarations. sql/share/errmsg.txt: If someone tampered with mysql.event table after the server has started we no longer give him/her a complete report what was actually broken. Do not send the user to look at the error log in such case, as there is nothing there (check_table_intact is not executed). sql/sp_head.cc: Update to a new declaration of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_db.cc: New invocation scheme of Events module. sql/sql_parse.cc: Move more logic to Events module. Make sure that we are consistent in the way access rights are checked for Events DDL: always after committing the current transaction and checking the system tables. sql/sql_show.cc: Update to the new declarations of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_test.cc: New invocation scheme of events. sql/table.cc: mysql.event is a system table. Update check_table_intact to be concurrent, more verbose, and less smart. sql/table.h: Add a helper method. mysql-test/r/events_trans.result: New BitKeeper file ``mysql-test/r/events_trans.result'' mysql-test/t/events_trans.test: New BitKeeper file ``mysql-test/t/events_trans.test'': test cases for Event Scheduler that require a transactional storage engine.
2007-04-05 13:24:34 +02:00
db_repository= db_repository_arg;
Fix for bug#22740 Events: Decouple Event_queue from Event_db_repository This patch implements the idea of the bug report by making Event_queue unaware of Event_db_repository by making a higher level class - Events, which is aware of most of all classes, responsible for passing all data needed for adding/updating/deleting an event to/from the queue. Introduces few new classes : - Event_worker_thread - Event_queue_element_for_exec sql/event_data_objects.cc: Introduced a new class Event_queue_element_for_exec According to Konstantin it should be named Event_name and hold only two LEX_STRINGs but `dropped` is not saved on disk and will require additional logic in Event_worker_thread class, after loading to compute whether the event should be dropped or not. It's easier just to pass this flag around. Removed Event_queue_element::drop(). This method was a source of a race condition. At the place where the event should be dropped we call Events::drop_event() which is the only code-flow for dropping. In addition, because ::drop_event() holds Events::LOCK_metadata there is no source of race conditions. Before this patch dropping from ::drop() wasn't under LOCK_metadata and races were possible. Because Events::open_event_table was removed as a method, provisionally events_event_db_repository was exported from events.cc till a solution is build where Event_queue_element does not access directly mysql.event. sql/event_data_objects.h: New class Event_queue_element_for_exec added which is returned from Event_queue::get_top_if_time() and passed through Event_scheduler to Event_worker_thread. There by using the (db)name Event_job_data is instanciated and executed. Dropped Event_queue_element::drop() thd was moved out of Event_job_data as it is now part of Event_queue_element_for_exec sql/event_queue.cc: Removed dependency of Event_queue on Event_db_repository. The instantiation of Event_job_data was moved to class Event_worker_thread In place is a return of an object of Event_queue_element_for_exec is used later for instantiating Event_job_data. The `dropped` flag of Event_queue_element is passed over Event_queue_element_for_exec to the code in Event_worker_thread. sql/event_queue.h: Removed dependency of Event_queue on Event_db_repository Removed dependency on Event_scheduler sql/event_scheduler.cc: Added class Event_worker_thread, which is used during the execution of an event. It has a static init() method to get a pointer to Event_db_repository to be used for instantiation of Event_job_data object. This object it then executed. sql/event_scheduler.h: Added class Event_worker_thread, which is used during the execution of an event. sql/events.cc: Removed Events::open_event_table() because it was a product of a bad architecture. sql/events.h: Removed friend definition, unneeded. Fixed Events::drop_event() to have the previous signature without bool only_from_disk sql/sql_parse.cc: Fix call
2007-01-29 18:46:29 +01:00
}
void
run(THD *thd, Event_queue_element_for_exec *event);
private:
void
print_warnings(THD *thd, Event_job_data *et);
static Event_db_repository *db_repository;
};
class Event_scheduler
{
public:
A set of changes aiming to make the Event Scheduler more user-friendly when there are no up-to-date system tables to support it: - initialize the scheduler before reporting "Ready for connections". This ensures that warnings, if any, are printed before "Ready for connections", and this message is not mangled. - do not abort the scheduler if there are no system tables - check the tables once at start up, remember the status and disable the scheduler if the tables are not up to date. If one attempts to use the scheduler with bad tables, issue an error message. - clean up the behaviour of the module under LOCK TABLES and pre-locking mode - make sure implicit commit of Events DDL works as expected. - add more tests Collateral clean ups in the events code. This patch fixes Bug#23631 Events: SHOW VARIABLES doesn't work when mysql.event is damaged mysql-test/r/events.result: Update results. mysql-test/r/events_bugs.result: Update results. mysql-test/r/events_restart_phase1.result: Update results. mysql-test/r/events_restart_phase2.result: Update results. mysql-test/r/events_restart_phase3.result: Update results. mysql-test/r/events_scheduling.result: Update results. mysql-test/r/events_time_zone.result: Update results. mysql-test/t/events.test: Add new tests for tampering with mysql.event and some more tests for sub-statements, LOCK TABLES mode and pre-locking. mysql-test/t/events_bugs.test: Move the non-concurrent part of test for Bug 16420 to this file. mysql-test/t/events_restart_phase1.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase2.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_restart_phase3.test: Rewrite events_restart_* tests to take into account that now we check mysql.event table only once, at server startup. mysql-test/t/events_scheduling.test: Add more coverage for event_scheduler global variable. mysql-test/t/events_time_zone.test: Move the non-concurrent part of the tests for Bug 16420 to events_bugs.test sql/event_data_objects.cc: Move update_timing_fields functionality to Event_db_repository. Make loading of events from a table record more robust to tampering with the table - now we do not check mysql.event on every table open. sql/event_data_objects.h: Cleanup. sql/event_db_repository.cc: Now Event_db_repository is responsible for table I/O only. All the logic of events DDL is handled outside, in Events class please refer to the added test coverage to see how this change affected the behavior of Event Scheduler. Dependency on sp_head.h and sp.h removed. Make this module robust to tweaks with mysql.event table. Move check_system_tables from events.cc to this file sql/event_db_repository.h: Cleanup declarations (remove unused ones, change return type to bool from int). sql/event_queue.cc: Update to adapt to the new start up scheme of the Event Scheduler. sql/event_queue.h: Cleanup declarations. sql/event_scheduler.cc: Make all the error messages uniform: [SEVERITY] Event Scheduler: [user][schema.event] message Using append_identifier for error logging was an overkill - we may need it only if the system character set may have NUL (null character) as part of a valid identifier, this is currently never the case, whereas additional quoting did not look nice in the log. sql/event_scheduler.h: Cleanup the headers. sql/events.cc: Use a different start up procedure of Event Scheduler: - at start up, try to check the system tables first. If they are not up-to-date, disable the scheduler. - try to load all the active events. In case of a load error, abort start up. - do not parse an event on start up. Parsing only gives some information about event validity, but far not all. Consolidate the business logic of Events DDL in this module. Now opt_event_scheduler may change after start up and thus is protected by LOCK_event_metadata mutex. sql/events.h: Use all-static-data-members approach to implement Singleton pattern. sql/mysqld.cc: New invocation scheme of Events. Move some logic to events.cc. Initialize the scheduler before reporting "Ready for connections". sql/set_var.cc: Clean up sys_var_thd_sql_mode::symbolic_mode_representation to work with a LEX_STRING. Move more logic related to @@events_scheduler global variable to Events module. sql/set_var.h: Update declarations. sql/share/errmsg.txt: If someone tampered with mysql.event table after the server has started we no longer give him/her a complete report what was actually broken. Do not send the user to look at the error log in such case, as there is nothing there (check_table_intact is not executed). sql/sp_head.cc: Update to a new declaration of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_db.cc: New invocation scheme of Events module. sql/sql_parse.cc: Move more logic to Events module. Make sure that we are consistent in the way access rights are checked for Events DDL: always after committing the current transaction and checking the system tables. sql/sql_show.cc: Update to the new declarations of sys_var_thd_sql_mode::symbolic_mode_representation sql/sql_test.cc: New invocation scheme of events. sql/table.cc: mysql.event is a system table. Update check_table_intact to be concurrent, more verbose, and less smart. sql/table.h: Add a helper method. mysql-test/r/events_trans.result: New BitKeeper file ``mysql-test/r/events_trans.result'' mysql-test/t/events_trans.test: New BitKeeper file ``mysql-test/t/events_trans.test'': test cases for Event Scheduler that require a transactional storage engine.
2007-04-05 13:24:34 +02:00
Event_scheduler(Event_queue *event_queue_arg);
~Event_scheduler();
/* State changing methods follow */
bool
start();
bool
stop();
/*
WL#3337 (Event scheduler new architecture) Post-review fixes. Mostly whitespace, int-to-bool return value, fixed comments sql/Makefile.am: compile all submodules of Events before compiling the facade sql/event_data_objects.cc: - Use initialization list - Clean whitespaces - Shorten comments - Fix comments sql/event_data_objects.h: - Fix whitespace sql/event_db_repository.cc: - Change return type from int to bool where only one error code is returned. - Don't use macros but get the maximal number of characters in a column from the column - Fix comments - Make functions which has return value but it's not used - void. sql/event_db_repository.h: - Methods with only one error code int -> bool return value - Remove declaration of fill_schema_events, a function that does not exist sql/event_queue.cc: - Use initialization lists - Let find_n_remove_event delete the object thus making the code more robust. The caller could forget to destruct the object. In addition, find_n_remove_element() does not return a value. - Move check_system_tables() to class Events - Fix comments sql/event_queue.h: - Whitespace changes - init_queue() should allow passing of THD - check_system_tables moved to class Events - find_n_remove_event() is now void sql/event_scheduler.cc: - Initialize res before use - Remove end stop from message sql/event_scheduler.h: Add uninitialized state. The scheduler is in it before init_scheduler() is called. The rationale is that otherwise state has no value before the call. If the system tables were damaged the scheduler won't be initialized but in Events::deinit() Event_scheduler::stop() will be called and this will touch state, generating valgrind warning at minimum. sql/events.cc: - Whitespace changes - Fix comments - Make methods which have only one error code be bool instead of int - Create temporarily a THD to be used for the initialization of Event_queue - Event_queue::check_system_tables() moved to Events::check_system_tables - is_started() is renamed to is_execution_of_events_started() sql/events.h: - Whitespace changes - When a method returns only one error code it should be bool, not int - is_started() renamed to is_execution_of_events_started() sql/set_var.cc: is_started() is renamed to is_execution_of_events_started() sql/sql_db.cc: The return code is not used, thus don't return anything and drop_schema_events() is now void. sql/sql_yacc.yy: - Fix comments - Remove unneeded initialization which is performed in lex_init() sql/share/errmsg.txt: New error message sql/table.cc: - Fix comments - make table_check_intact() accespt const *table_def sql/table.h: Make table_check_intact() accespt const *table_def
2006-08-17 14:22:59 +02:00
Need to be public because has to be called from the function
passed to pthread_create.
*/
bool
run(THD *thd);
WL#3337 (Event scheduler new architecture) Post-review fixes. Mostly whitespace, int-to-bool return value, fixed comments sql/Makefile.am: compile all submodules of Events before compiling the facade sql/event_data_objects.cc: - Use initialization list - Clean whitespaces - Shorten comments - Fix comments sql/event_data_objects.h: - Fix whitespace sql/event_db_repository.cc: - Change return type from int to bool where only one error code is returned. - Don't use macros but get the maximal number of characters in a column from the column - Fix comments - Make functions which has return value but it's not used - void. sql/event_db_repository.h: - Methods with only one error code int -> bool return value - Remove declaration of fill_schema_events, a function that does not exist sql/event_queue.cc: - Use initialization lists - Let find_n_remove_event delete the object thus making the code more robust. The caller could forget to destruct the object. In addition, find_n_remove_element() does not return a value. - Move check_system_tables() to class Events - Fix comments sql/event_queue.h: - Whitespace changes - init_queue() should allow passing of THD - check_system_tables moved to class Events - find_n_remove_event() is now void sql/event_scheduler.cc: - Initialize res before use - Remove end stop from message sql/event_scheduler.h: Add uninitialized state. The scheduler is in it before init_scheduler() is called. The rationale is that otherwise state has no value before the call. If the system tables were damaged the scheduler won't be initialized but in Events::deinit() Event_scheduler::stop() will be called and this will touch state, generating valgrind warning at minimum. sql/events.cc: - Whitespace changes - Fix comments - Make methods which have only one error code be bool instead of int - Create temporarily a THD to be used for the initialization of Event_queue - Event_queue::check_system_tables() moved to Events::check_system_tables - is_started() is renamed to is_execution_of_events_started() sql/events.h: - Whitespace changes - When a method returns only one error code it should be bool, not int - is_started() renamed to is_execution_of_events_started() sql/set_var.cc: is_started() is renamed to is_execution_of_events_started() sql/sql_db.cc: The return code is not used, thus don't return anything and drop_schema_events() is now void. sql/sql_yacc.yy: - Fix comments - Remove unneeded initialization which is performed in lex_init() sql/share/errmsg.txt: New error message sql/table.cc: - Fix comments - make table_check_intact() accespt const *table_def sql/table.h: Make table_check_intact() accespt const *table_def
2006-08-17 14:22:59 +02:00
/* Information retrieving methods follow */
WL#3337 (Event scheduler new architecture) This patch makes the relationship between Event_scheduler and Event_queue unidirectional from the former to the latter. The change is that the conditional on which the scheduler sleeped has been moved to the Event_queue and the latter does not call anymore Event_scheduler::queue_changed(), which in turn has be removed. sql/event_queue.cc: Remove dependency of Event_queue on Event_scheduler but not vice versa. Event_scheduler polls whether there is time to execute an event. Removed notify_observers() as the way of calling has changed. Added Event_queue::cond_wait() similar to Event_scheduler::cond_wait(). sql/event_queue.h: init_queue() does not need anymore Event_scheduler object because the relationship is now one-way. Event_scheduler knows about Event_queue but not vice versa. get_top_execution_if_time() does by itself the waiting instead of returning abstime. This simplifies the code in Event_scheduler::run() get_top_execution_if_time() returns only if job_data != NULL or if the scheduler thread was killed. notify_observers() is no more used and therefore removed. Added Event_queue::cond_wait() because now there is waiting on a conditional variable in Event_queue too (like in Event_scheduler for ::stop()). sql/event_scheduler.cc: Change the relationship between Event_scheduler & Event_queue. Event_queue does not know anymore about Event_scheduler. When the scheduler calls get_top_element_if_time() it may fall asleep on a conditional of Event_queue, if either the queue is empty or it's still not time for activation. When the method returns it will return a non-null address, namely an object to be executed. If the return value is NULL, the thread was killed by a call to Event_scheduler::stop() (we assert this). sql/event_scheduler.h: Remove queue_changed() as it is obsoleted by making the relationship between Event_scheduler and Event_queue one-way, from the former to the latter. Event_queue now does not know about Event_scheduler. get_state() is changed to is_running(). The state enum should be private, as it is not needed to be seen from outside anymore. sql/events.cc: Event_queue does not need anymore a pointer to Event_scheduler.
2006-08-31 17:18:39 +02:00
bool
is_running();
void
dump_internal_status();
private:
uint
workers_count();
/* helper functions */
bool
execute_top(Event_queue_element_for_exec *event_name);
/* helper functions for working with mutexes & conditionals */
void
lock_data(const char *func, uint line);
void
unlock_data(const char *func, uint line);
void
WL#3337 (Events new architecture) Small updates before patch submit. client/mysqltest.c: allow --valgrind option to mysqltest so one can be able to detect whether the test is running under valgrind by having $VALGRIND_TEST, similar to BIG_TEST, in the test file. mysql-test/mysql-test-run.pl: If the test suite is running under valgrind start mysqltest with --valgrind to inform that we run valgrind. mysqltest will provide $VALGRIND_TEST for the test cases. mysql-test/r/events_bugs.result: update result mysql-test/r/events_scheduling.result: update result mysql-test/t/events.test: Increase times or the test will fail under valgrind mysql-test/t/events_bugs.test: Increase times or the test will fail under valgrind mysql-test/t/events_scheduling.test: Remove faulty test Disable the test case for valgrind sql/event_data_objects.cc: count the number of executions sql/event_data_objects.h: flags is not used at all add execution_count to count the number of executions sql/event_db_repository.cc: Initialize wherever needed. Add a comment regarding valgrind warning. sql/event_queue.cc: more debug info in the trace log sql/event_scheduler.cc: Use macro COND_STATE_WAIT() in all cases we need waiting on condition. Hence, we can trace locking, attemption to lock and more with SHOW SCHEDULER STATUS sql/event_scheduler.h: Change the declaration of cond_wait to accept THD sql/events.cc: fix memory leak. Destroy event_queue mysql-test/include/not_valgrind.inc: New BitKeeper file ``mysql-test/include/not_valgrind.inc'' mysql-test/r/not_valgrind.require: New BitKeeper file ``mysql-test/r/not_valgrind.require''
2006-07-17 16:52:45 +02:00
cond_wait(THD *thd, struct timespec *abstime, const char* msg,
const char *func, uint line);
pthread_mutex_t LOCK_scheduler_state;
WL#3337 (Event scheduler new architecture) This patch makes the relationship between Event_scheduler and Event_queue unidirectional from the former to the latter. The change is that the conditional on which the scheduler sleeped has been moved to the Event_queue and the latter does not call anymore Event_scheduler::queue_changed(), which in turn has be removed. sql/event_queue.cc: Remove dependency of Event_queue on Event_scheduler but not vice versa. Event_scheduler polls whether there is time to execute an event. Removed notify_observers() as the way of calling has changed. Added Event_queue::cond_wait() similar to Event_scheduler::cond_wait(). sql/event_queue.h: init_queue() does not need anymore Event_scheduler object because the relationship is now one-way. Event_scheduler knows about Event_queue but not vice versa. get_top_execution_if_time() does by itself the waiting instead of returning abstime. This simplifies the code in Event_scheduler::run() get_top_execution_if_time() returns only if job_data != NULL or if the scheduler thread was killed. notify_observers() is no more used and therefore removed. Added Event_queue::cond_wait() because now there is waiting on a conditional variable in Event_queue too (like in Event_scheduler for ::stop()). sql/event_scheduler.cc: Change the relationship between Event_scheduler & Event_queue. Event_queue does not know anymore about Event_scheduler. When the scheduler calls get_top_element_if_time() it may fall asleep on a conditional of Event_queue, if either the queue is empty or it's still not time for activation. When the method returns it will return a non-null address, namely an object to be executed. If the return value is NULL, the thread was killed by a call to Event_scheduler::stop() (we assert this). sql/event_scheduler.h: Remove queue_changed() as it is obsoleted by making the relationship between Event_scheduler and Event_queue one-way, from the former to the latter. Event_queue now does not know about Event_scheduler. get_state() is changed to is_running(). The state enum should be private, as it is not needed to be seen from outside anymore. sql/events.cc: Event_queue does not need anymore a pointer to Event_scheduler.
2006-08-31 17:18:39 +02:00
enum enum_state
{
INITIALIZED = 0,
WL#3337 (Event scheduler new architecture) This patch makes the relationship between Event_scheduler and Event_queue unidirectional from the former to the latter. The change is that the conditional on which the scheduler sleeped has been moved to the Event_queue and the latter does not call anymore Event_scheduler::queue_changed(), which in turn has be removed. sql/event_queue.cc: Remove dependency of Event_queue on Event_scheduler but not vice versa. Event_scheduler polls whether there is time to execute an event. Removed notify_observers() as the way of calling has changed. Added Event_queue::cond_wait() similar to Event_scheduler::cond_wait(). sql/event_queue.h: init_queue() does not need anymore Event_scheduler object because the relationship is now one-way. Event_scheduler knows about Event_queue but not vice versa. get_top_execution_if_time() does by itself the waiting instead of returning abstime. This simplifies the code in Event_scheduler::run() get_top_execution_if_time() returns only if job_data != NULL or if the scheduler thread was killed. notify_observers() is no more used and therefore removed. Added Event_queue::cond_wait() because now there is waiting on a conditional variable in Event_queue too (like in Event_scheduler for ::stop()). sql/event_scheduler.cc: Change the relationship between Event_scheduler & Event_queue. Event_queue does not know anymore about Event_scheduler. When the scheduler calls get_top_element_if_time() it may fall asleep on a conditional of Event_queue, if either the queue is empty or it's still not time for activation. When the method returns it will return a non-null address, namely an object to be executed. If the return value is NULL, the thread was killed by a call to Event_scheduler::stop() (we assert this). sql/event_scheduler.h: Remove queue_changed() as it is obsoleted by making the relationship between Event_scheduler and Event_queue one-way, from the former to the latter. Event_queue now does not know about Event_scheduler. get_state() is changed to is_running(). The state enum should be private, as it is not needed to be seen from outside anymore. sql/events.cc: Event_queue does not need anymore a pointer to Event_scheduler.
2006-08-31 17:18:39 +02:00
RUNNING,
STOPPING
};
/* This is the current status of the life-cycle of the scheduler. */
enum enum_state state;
WL#3337 (Event scheduler new architecture) This patch makes the relationship between Event_scheduler and Event_queue unidirectional from the former to the latter. The change is that the conditional on which the scheduler sleeped has been moved to the Event_queue and the latter does not call anymore Event_scheduler::queue_changed(), which in turn has be removed. sql/event_queue.cc: Remove dependency of Event_queue on Event_scheduler but not vice versa. Event_scheduler polls whether there is time to execute an event. Removed notify_observers() as the way of calling has changed. Added Event_queue::cond_wait() similar to Event_scheduler::cond_wait(). sql/event_queue.h: init_queue() does not need anymore Event_scheduler object because the relationship is now one-way. Event_scheduler knows about Event_queue but not vice versa. get_top_execution_if_time() does by itself the waiting instead of returning abstime. This simplifies the code in Event_scheduler::run() get_top_execution_if_time() returns only if job_data != NULL or if the scheduler thread was killed. notify_observers() is no more used and therefore removed. Added Event_queue::cond_wait() because now there is waiting on a conditional variable in Event_queue too (like in Event_scheduler for ::stop()). sql/event_scheduler.cc: Change the relationship between Event_scheduler & Event_queue. Event_queue does not know anymore about Event_scheduler. When the scheduler calls get_top_element_if_time() it may fall asleep on a conditional of Event_queue, if either the queue is empty or it's still not time for activation. When the method returns it will return a non-null address, namely an object to be executed. If the return value is NULL, the thread was killed by a call to Event_scheduler::stop() (we assert this). sql/event_scheduler.h: Remove queue_changed() as it is obsoleted by making the relationship between Event_scheduler and Event_queue one-way, from the former to the latter. Event_queue now does not know about Event_scheduler. get_state() is changed to is_running(). The state enum should be private, as it is not needed to be seen from outside anymore. sql/events.cc: Event_queue does not need anymore a pointer to Event_scheduler.
2006-08-31 17:18:39 +02:00
THD *scheduler_thd;
pthread_cond_t COND_state;
WL#3337 (Event scheduler new architecture) Post-review fixes. Mostly whitespace, int-to-bool return value, fixed comments sql/Makefile.am: compile all submodules of Events before compiling the facade sql/event_data_objects.cc: - Use initialization list - Clean whitespaces - Shorten comments - Fix comments sql/event_data_objects.h: - Fix whitespace sql/event_db_repository.cc: - Change return type from int to bool where only one error code is returned. - Don't use macros but get the maximal number of characters in a column from the column - Fix comments - Make functions which has return value but it's not used - void. sql/event_db_repository.h: - Methods with only one error code int -> bool return value - Remove declaration of fill_schema_events, a function that does not exist sql/event_queue.cc: - Use initialization lists - Let find_n_remove_event delete the object thus making the code more robust. The caller could forget to destruct the object. In addition, find_n_remove_element() does not return a value. - Move check_system_tables() to class Events - Fix comments sql/event_queue.h: - Whitespace changes - init_queue() should allow passing of THD - check_system_tables moved to class Events - find_n_remove_event() is now void sql/event_scheduler.cc: - Initialize res before use - Remove end stop from message sql/event_scheduler.h: Add uninitialized state. The scheduler is in it before init_scheduler() is called. The rationale is that otherwise state has no value before the call. If the system tables were damaged the scheduler won't be initialized but in Events::deinit() Event_scheduler::stop() will be called and this will touch state, generating valgrind warning at minimum. sql/events.cc: - Whitespace changes - Fix comments - Make methods which have only one error code be bool instead of int - Create temporarily a THD to be used for the initialization of Event_queue - Event_queue::check_system_tables() moved to Events::check_system_tables - is_started() is renamed to is_execution_of_events_started() sql/events.h: - Whitespace changes - When a method returns only one error code it should be bool, not int - is_started() renamed to is_execution_of_events_started() sql/set_var.cc: is_started() is renamed to is_execution_of_events_started() sql/sql_db.cc: The return code is not used, thus don't return anything and drop_schema_events() is now void. sql/sql_yacc.yy: - Fix comments - Remove unneeded initialization which is performed in lex_init() sql/share/errmsg.txt: New error message sql/table.cc: - Fix comments - make table_check_intact() accespt const *table_def sql/table.h: Make table_check_intact() accespt const *table_def
2006-08-17 14:22:59 +02:00
Event_queue *queue;
uint mutex_last_locked_at_line;
uint mutex_last_unlocked_at_line;
const char* mutex_last_locked_in_func;
const char* mutex_last_unlocked_in_func;
bool mutex_scheduler_data_locked;
bool waiting_on_cond;
WL#3337 (Event scheduler new architecture) Post-review fixes. Mostly whitespace, int-to-bool return value, fixed comments sql/Makefile.am: compile all submodules of Events before compiling the facade sql/event_data_objects.cc: - Use initialization list - Clean whitespaces - Shorten comments - Fix comments sql/event_data_objects.h: - Fix whitespace sql/event_db_repository.cc: - Change return type from int to bool where only one error code is returned. - Don't use macros but get the maximal number of characters in a column from the column - Fix comments - Make functions which has return value but it's not used - void. sql/event_db_repository.h: - Methods with only one error code int -> bool return value - Remove declaration of fill_schema_events, a function that does not exist sql/event_queue.cc: - Use initialization lists - Let find_n_remove_event delete the object thus making the code more robust. The caller could forget to destruct the object. In addition, find_n_remove_element() does not return a value. - Move check_system_tables() to class Events - Fix comments sql/event_queue.h: - Whitespace changes - init_queue() should allow passing of THD - check_system_tables moved to class Events - find_n_remove_event() is now void sql/event_scheduler.cc: - Initialize res before use - Remove end stop from message sql/event_scheduler.h: Add uninitialized state. The scheduler is in it before init_scheduler() is called. The rationale is that otherwise state has no value before the call. If the system tables were damaged the scheduler won't be initialized but in Events::deinit() Event_scheduler::stop() will be called and this will touch state, generating valgrind warning at minimum. sql/events.cc: - Whitespace changes - Fix comments - Make methods which have only one error code be bool instead of int - Create temporarily a THD to be used for the initialization of Event_queue - Event_queue::check_system_tables() moved to Events::check_system_tables - is_started() is renamed to is_execution_of_events_started() sql/events.h: - Whitespace changes - When a method returns only one error code it should be bool, not int - is_started() renamed to is_execution_of_events_started() sql/set_var.cc: is_started() is renamed to is_execution_of_events_started() sql/sql_db.cc: The return code is not used, thus don't return anything and drop_schema_events() is now void. sql/sql_yacc.yy: - Fix comments - Remove unneeded initialization which is performed in lex_init() sql/share/errmsg.txt: New error message sql/table.cc: - Fix comments - make table_check_intact() accespt const *table_def sql/table.h: Make table_check_intact() accespt const *table_def
2006-08-17 14:22:59 +02:00
ulonglong started_events;
private:
/* Prevent use of these */
Event_scheduler(const Event_scheduler &);
void operator=(Event_scheduler &);
};
/**
@} (End of group Event_Scheduler)
*/
fix for bug #17619 Scheduler race conditions - Scheduler is either initialized at server start or never. Starting & stopping is now suspending & resuming. - The scheduler has clear OO interface - Now all calls to the scheduler are synchronous - GLOBAL event_scheduler uses thd::sys_var_tmp (see set_var.cc) - External API is encapsulated into class Events - Includes fixes for all comments of Kostja's review of 19.05.2005 Starting to merge into 5.1-release (5.1.10) and push BitKeeper/etc/ignore: Added libmysqld/event_scheduler.cc to the ignore list libmysqld/Makefile.am: executor -> scheduler mysql-test/r/events.result: update result mysql-test/r/events_bugs.result: update result mysql-test/r/events_logs_tests.result: update result mysql-test/r/events_microsec.result: update result mysql-test/r/events_scheduling.result: update result mysql-test/r/events_stress.result: update result mysql-test/t/disabled.def: enable these tests mysql-test/t/events.test: optimize the test a bit for speed, save some seconds runtime remove FULL from SHOW EVENTS mostly use I_S.EVENTS mysql-test/t/events_bugs.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_logs_tests.test: optimize the test a bit for speed mysql-test/t/events_microsec.test: Skip irrelevant for the current design tests - all events are loaded on server startup. Change in mysql.event will be visible on next server start. Don't use numeric error codes. mysql-test/t/events_scheduling.test: broader test mysql-test/t/events_stress.test: Rework the test to the new architecture of suspending/resuming. Use less events, no need for thousands, hundreds is still ok. sql/Makefile.am: executor -> scheduler sql/cmakelists.txt: executor -> scheduler sql/event.cc: - remove todo comments - remove unneded evex_queue abstraction functions - move events_init() and events_shutdown() from event_executor.cc to here - export db_create_event - remove evex_load_and_compile_event, part of class Event_scheduler - integrate the public interface found in event.h and used by sql_parse.cc to use the new class Event_scheduler. sql/event.h: - add COND_finished so if one thread kills a running event it waits on this - export callback event_timed_definer_equal, event_timed_identifier_equal(), event_timed_name_equal and event_timed_db_equal() to be used by Event_scheduler::drop_matching_events() - cleanup event.h - encapsulated all external interface into class Events sql/event_executor.cc: make it empty, will delete after that sql/event_priv.h: - more things in the private header - remove event queue abstraction functions. tightly bind to QUEUE - export privately db_drop_event, db_find_event, db_create_event() - made change_security_context() and restore_security_context() free functions sql/event_timed.cc: - fix calculation of time when ENDS is set (STARTS is always set) - during Event_timed::compile() set the right Security_ctx. Prevents a crash during Event_scheduler::load_events_from_db() - add Event_timed::kill_thread() - implement event_timed_*_equal() - made change_security_context() and restore_security_context() free functions. - Comments cleanups sql/lex.h: new word scheduler for SHOW SCHEDULER STATUS (available only debug builds) sql/log.cc: move these from event_scheduler.cc sql/mysql_priv.h: refactor kill_one_thread export sql_print_message_func and sql_print_message_handlers sql/mysqld.cc: In close_connections, called by kill_server() skip the main scheduler thread and use events_shutdown() for shutting down the scheduler, in the same manner it's done for RPL. Add a new value to --event-scheduler : 0 <- No scheduler available 1 <- Start with scheduler enabled 2 <- Start with scheduler suspended sql/repl_failsafe.cc: refactor thd::system_thread to be an enum sql/set_var.cc: move sys_var_event_executor::update() to set_var.cc executor -> scheduler use thd::sys_var_tmp sql/set_var.h: executor -> scheduler sql/share/errmsg.txt: 3 new error messages sql/sql_class.cc: refactor thd::system_thread to be an enum . more type-safety sql/sql_class.h: refactor thd::system_thread to be an enum . more type-safety sql/sql_db.cc: get the error from evex_drop_schema_events sql/sql_error.h: export warning_level_names sql/sql_lex.h: new command SHOW SCHEDULER STATUS, available only in debug build and for debug purposes. sql/sql_parse.cc: refactor kill_one_thread() -> does the *dirty* work, and sql_kill just the reporting. add handler for SQLCOM_SHOW_SCHEDULER_STATUS sql/sql_show.cc: fix verbosity handling (this will be obsoleted anyway by the fix for 17394). sql/sql_yacc.yy: remove FULL from SHOW EVENTS add SHOW SCHEDULER STATUS in debug builds sql/table.cc: Fix valgrind warning.
2006-05-22 20:46:13 +02:00
#endif /* _EVENT_SCHEDULER_H_ */