Transaction

TransactionId 是一個32位的無符號整數,事務和子事務的id,全是按照一個順序來的,當這個32位整數用完一圈後,就需要一些特殊的操作

/*----------
	 * Check to see if it's safe to assign another XID.  This protects against
	 * catastrophic data loss due to XID wraparound.  The basic rules are:
	 *
	 * If we're past xidVacLimit, start trying to force autovacuum cycles.
	 * If we're past xidWarnLimit, start issuing warnings.
	   To avoid a database shutdown, execute a database-wide VACUUM in that database.
	 * If we're past xidStopLimit, refuse to execute transactions, unless
	 * we are running in a standalone backend (which gives an escape hatch
	 * to the DBA who somehow got past the earlier defenses).
           Stop the postmaster and use a standalone backend to vacuum that database.
	 *----------
	 */

 

 

 

 

/* The transaction handle; every session has a trx object which is freed only when the session is freed; in addition there may be session-less transactions rolling back after a database recovery */ struct trx_struct{ ulint magic_n; /* These fields are not protected by any mutex. */ const char* op_info; /*!< English text describing the current operation, or an empty string */ ulint conc_state; /*!< state of the trx from the point of view of concurrency control: TRX_ACTIVE, TRX_COMMITTED_IN_MEMORY, ... */ /*------------------------------*/ /* MySQL has a transaction coordinator to coordinate two phase commit between multiple storage engines and the binary log. When an engine participates in a transaction, it's responsible for registering itself using the trans_register_ha() API. */ unsigned is_registered:1;/* This flag is set to 1 after the transaction has been registered with the coordinator using the XA API, and is set to 0 after commit or rollback. */ unsigned owns_prepare_mutex:1;/* 1 if owns prepare mutex, if this is set to 1 then registered should also be set to 1. This is used in the XA code */ /*------------------------------*/ ulint isolation_level;/* TRX_ISO_REPEATABLE_READ, ... */ ulint check_foreigns; /* normally TRUE, but if the user wants to suppress foreign key checks, (in table imports, for example) we set this FALSE */ ulint check_unique_secondary; /* normally TRUE, but if the user wants to speed up inserts by suppressing unique key checks for secondary indexes when we decide if we can use the insert buffer for them, we set this FALSE */ ulint support_xa; /*!< normally we do the XA two-phase commit steps, but by setting this to FALSE, one can save CPU time and about 150 bytes in the undo log size as then we skip XA steps */ ulint flush_log_later;/* In 2PC, we hold the prepare_commit mutex across both phases. In that case, we defer flush of the logs to disk until after we release the mutex. */ ulint must_flush_log_later;/* this flag is set to TRUE in trx_commit_off_kernel() if flush_log_later was TRUE, and there were modifications by the transaction; in that case we must flush the log in trx_commit_complete_for_mysql() */ ulint duplicates; /*!< TRX_DUP_IGNORE | TRX_DUP_REPLACE */ ulint has_search_latch; /* TRUE if this trx has latched the search system latch in S-mode */ ulint deadlock_mark; /*!< a mark field used in deadlock checking algorithm. */ trx_dict_op_t dict_operation; /**< @see enum trx_dict_op */ /* Fields protected by the srv_conc_mutex. */ ulint declared_to_be_inside_innodb; /* this is TRUE if we have declared this transaction in srv_conc_enter_innodb to be inside the InnoDB engine */ /* Fields protected by dict_operation_lock. The very latch it is used to track. */ ulint dict_operation_lock_mode; /*!< 0, RW_S_LATCH, or RW_X_LATCH: the latch mode trx currently holds on dict_operation_lock */ /* All the next fields are protected by the kernel mutex, except the undo logs which are protected by undo_mutex */ ulint is_purge; /*!< 0=user transaction, 1=purge */ ulint is_recovered; /*!< 0=normal transaction, 1=recovered, must be rolled back */ ulint que_state; /*!< valid when conc_state == TRX_ACTIVE: TRX_QUE_RUNNING, TRX_QUE_LOCK_WAIT, ... */ ulint handling_signals;/* this is TRUE as long as the trx is handling signals */ time_t start_time; /*!< time the trx object was created or the state last time became TRX_ACTIVE */ trx_id_t id; /*!< transaction id */ XID xid; /*!< X/Open XA transaction identification to identify a transaction branch */ trx_id_t no; /*!< transaction serialization number == max trx id when the transaction is moved to COMMITTED_IN_MEMORY state */ ib_uint64_t commit_lsn; /*!< lsn at the time of the commit */ table_id_t table_id; /*!< Table to drop iff dict_operation is TRUE, or 0. */ /*------------------------------*/ void* mysql_thd; /*!< MySQL thread handle corresponding to this trx, or NULL */ const char* mysql_log_file_name; /* if MySQL binlog is used, this field contains a pointer to the latest file name; this is NULL if binlog is not used */ ib_int64_t mysql_log_offset;/* if MySQL binlog is used, this field contains the end offset of the binlog entry */ os_thread_id_t mysql_thread_id;/* id of the MySQL thread associated with this transaction object */ ulint mysql_process_no;/* since in Linux, 'top' reports process id's and not thread id's, we store the process number too */ /*------------------------------*/ ulint n_mysql_tables_in_use; /* number of Innobase tables used in the processing of the current SQL statement in MySQL */ ulint mysql_n_tables_locked; /* how many tables the current SQL statement uses, except those in consistent read */ ulint search_latch_timeout; /* If we notice that someone is waiting for our S-lock on the search latch to be released, we wait in row0sel.c for BTR_SEA_TIMEOUT new searches until we try to keep the search latch again over calls from MySQL; this is intended to reduce contention on the search latch */ /*------------------------------*/ ulint n_tickets_to_enter_innodb; /* this can be > 0 only when declared_to_... is TRUE; when we come to srv_conc_innodb_enter, if the value here is > 0, we decrement this by 1 */ /*------------------------------*/ UT_LIST_NODE_T(trx_t) trx_list; /*!< list of transactions */ UT_LIST_NODE_T(trx_t) mysql_trx_list; /*!< list of transactions created for MySQL */ /*------------------------------*/ ulint error_state; /*!< 0 if no error, otherwise error number; NOTE That ONLY the thread doing the transaction is allowed to set this field: this is NOT protected by the kernel mutex */ const dict_index_t*error_info; /*!< if the error number indicates a duplicate key error, a pointer to the problematic index is stored here */ ulint error_key_num; /*!< if the index creation fails to a duplicate key error, a mysql key number of that index is stored here */ sess_t* sess; /*!< session of the trx, NULL if none */ que_t* graph; /*!< query currently run in the session, or NULL if none; NOTE that the query belongs to the session, and it can survive over a transaction commit, if it is a stored procedure with a COMMIT WORK statement, for instance */ ulint n_active_thrs; /*!< number of active query threads */ que_t* graph_before_signal_handling; /* value of graph when signal handling for this trx started: this is used to return control to the original query graph for error processing */ trx_sig_t sig; /*!< one signal object can be allocated in this space, avoiding mem_alloc */ UT_LIST_BASE_NODE_T(trx_sig_t) signals; /*!< queue of processed or pending signals to the trx */ UT_LIST_BASE_NODE_T(trx_sig_t) reply_signals; /*!< list of signals sent by the query threads of this trx for which a thread is waiting for a reply; if this trx is killed, the reply requests in the list must be canceled */ /*------------------------------*/ lock_t* wait_lock; /*!< if trx execution state is TRX_QUE_LOCK_WAIT, this points to the lock request, otherwise this is NULL */ ibool was_chosen_as_deadlock_victim; /* when the transaction decides to wait for a lock, it sets this to FALSE; if another transaction chooses this transaction as a victim in deadlock resolution, it sets this to TRUE */ time_t wait_started; /*!< lock wait started at this time */ UT_LIST_BASE_NODE_T(que_thr_t) wait_thrs; /*!< query threads belonging to this trx that are in the QUE_THR_LOCK_WAIT state */ /*------------------------------*/ mem_heap_t* lock_heap; /*!< memory heap for the locks of the transaction */ UT_LIST_BASE_NODE_T(lock_t) trx_locks; /*!< locks reserved by the transaction */ /*------------------------------*/ mem_heap_t* global_read_view_heap; /* memory heap for the global read view */ read_view_t* global_read_view; /* consistent read view associated to a transaction or NULL */ read_view_t* read_view; /*!< consistent read view used in the transaction or NULL, this read view if defined can be normal read view associated to a transaction (i.e. same as global_read_view) or read view associated to a cursor */ /*------------------------------*/ UT_LIST_BASE_NODE_T(trx_named_savept_t) trx_savepoints; /*!< savepoints set with SAVEPOINT ..., oldest first */ /*------------------------------*/ mutex_t undo_mutex; /*!< mutex protecting the fields in this section (down to undo_no_arr), EXCEPT last_sql_stat_start, which can be accessed only when we know that there cannot be any activity in the undo logs! */ undo_no_t undo_no; /*!< next undo log record number to assign; since the undo log is private for a transaction, this is a simple ascending sequence with no gaps; thus it represents the number of modified/inserted rows in a transaction */ trx_savept_t last_sql_stat_start; /* undo_no when the last sql statement was started: in case of an error, trx is rolled back down to this undo number; see note at undo_mutex! */ trx_rseg_t* rseg; /*!< rollback segment assigned to the transaction, or NULL if not assigned yet */ trx_undo_t* insert_undo; /*!< pointer to the insert undo log, or NULL if no inserts performed yet */ trx_undo_t* update_undo; /*!< pointer to the update undo log, or NULL if no update performed yet */ undo_no_t roll_limit; /*!< least undo number to undo during a rollback */ ulint pages_undone; /*!< number of undo log pages undone since the last undo log truncation */ trx_undo_arr_t* undo_no_arr; /*!< array of undo numbers of undo log records which are currently processed by a rollback operation */ /*------------------------------*/ ulint n_autoinc_rows; /*!< no. of AUTO-INC rows required for an SQL statement. This is useful for multi-row INSERTs */ ib_vector_t* autoinc_locks; /* AUTOINC locks held by this transaction. Note that these are also in the lock list trx_locks. This vector needs to be freed explicitly when the trx_t instance is desrtoyed */ /*------------------------------*/ char detailed_error[256]; /*!< detailed error message for last error, or empty. */ }; /** The transaction system central memory data structure; protected by the kernel mutex */ struct trx_sys_struct{ trx_id_t max_trx_id; /*!< The smallest number not yet assigned as a transaction id or transaction number */ UT_LIST_BASE_NODE_T(trx_t) trx_list; /*!< List of active and committed in memory transactions, sorted on trx id, biggest first */ UT_LIST_BASE_NODE_T(trx_t) mysql_trx_list; /*!< List of transactions created for MySQL */ UT_LIST_BASE_NODE_T(trx_rseg_t) rseg_list; /*!< List of rollback segment objects */ trx_rseg_t* latest_rseg; /*!< Latest rollback segment in the round-robin assignment of rollback segments to transactions */ trx_rseg_t* rseg_array[TRX_SYS_N_RSEGS]; /*!< Pointer array to rollback segments; NULL if slot not in use */ ulint rseg_history_len;/*!< Length of the TRX_RSEG_HISTORY list (update undo logs for committed transactions), protected by rseg->mutex */ UT_LIST_BASE_NODE_T(read_view_t) view_list; /*!< List of read views sorted on trx no, biggest first */ }; //開始一個事務很簡單，分配rollback segment,ID，插入到系統鏈表裏面 ibool trx_start_low( /*==========*/ trx_t* trx, /*!< in: transaction */ ulint rseg_id)/*!< in: rollback segment id; if ULINT_UNDEFINED is passed, the system chooses the rollback segment automatically in a round-robin fashion */ { trx_rseg_t* rseg; ut_ad(mutex_own(&kernel_mutex)); ut_ad(trx->rseg == NULL); if (trx->is_purge) { trx->id = 0; trx->conc_state = TRX_ACTIVE; trx->start_time = time(NULL); return(TRUE); } ut_ad(trx->conc_state != TRX_ACTIVE); ut_a(rseg_id == ULINT_UNDEFINED); rseg = trx_assign_rseg(srv_rollback_segments); trx->id = trx_sys_get_new_trx_id(); /* The initial value for trx->no: IB_ULONGLONG_MAX is used in read_view_open_now: */ trx->no = IB_ULONGLONG_MAX; trx->rseg = rseg; trx->conc_state = TRX_ACTIVE; trx->start_time = time(NULL); UT_LIST_ADD_FIRST(trx_list, trx_sys->trx_list, trx); return(TRUE); }