mariadb/storage/innobase/include/page0page.h

1341 lines
46 KiB
C
Raw Normal View History

/*****************************************************************************
2016-06-21 14:21:03 +02:00
Copyright (c) 1994, 2016, Oracle and/or its affiliates. All Rights Reserved.
2017-09-13 09:27:15 +03:00
Copyright (c) 2013, 2017, MariaDB Corporation.
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 the Free Software
Foundation; version 2 of the License.
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.,
51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA
*****************************************************************************/
/**************************************************//**
@file include/page0page.h
Index page routines
Created 2/2/1994 Heikki Tuuri
*******************************************************/
#ifndef page0page_h
#define page0page_h
#include "univ.i"
#include "page0types.h"
#ifndef UNIV_INNOCHECKSUM
#include "fil0fil.h"
#include "buf0buf.h"
#include "data0data.h"
#include "dict0dict.h"
#include "rem0rec.h"
#endif /* !UNIV_INNOCHECKSUM*/
#include "fsp0fsp.h"
#ifndef UNIV_INNOCHECKSUM
#include "mtr0mtr.h"
#ifdef UNIV_MATERIALIZE
#undef UNIV_INLINE
#define UNIV_INLINE
#endif
/* PAGE HEADER
===========
Index page header starts at the first offset left free by the FIL-module */
typedef byte page_header_t;
#endif /* !UNIV_INNOCHECKSUM */
#define PAGE_HEADER FSEG_PAGE_DATA /* index page header starts at this
offset */
/*-----------------------------*/
#define PAGE_N_DIR_SLOTS 0 /* number of slots in page directory */
#define PAGE_HEAP_TOP 2 /* pointer to record heap top */
#define PAGE_N_HEAP 4 /* number of records in the heap,
bit 15=flag: new-style compact page format */
#define PAGE_FREE 6 /* pointer to start of page free record list */
#define PAGE_GARBAGE 8 /* number of bytes in deleted records */
#define PAGE_LAST_INSERT 10 /* pointer to the last inserted record, or
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
0 if this info has been reset by a delete,
for example */
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
/** This 10-bit field is usually 0. In B-tree index pages of
ROW_FORMAT=REDUNDANT tables, this byte can contain garbage if the .ibd
file was created in MySQL 4.1.0 or if the table resides in the system
tablespace and was created before MySQL 4.1.1 or MySQL 4.0.14.
In this case, the FIL_PAGE_TYPE would be FIL_PAGE_INDEX.
In ROW_FORMAT=COMPRESSED tables, this field is always 0, because
instant ADD COLUMN is not supported.
In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC tables, this field is
always 0, except in the root page of the clustered index after instant
ADD COLUMN.
Instant ADD COLUMN will change FIL_PAGE_TYPE to FIL_PAGE_TYPE_INSTANT
and initialize the PAGE_INSTANT field to the original number of
fields in the clustered index (dict_index_t::n_core_fields). The most
significant bits are in the first byte, and the least significant 5
bits are stored in the most significant 5 bits of PAGE_DIRECTION_B.
These FIL_PAGE_TYPE_INSTANT and PAGE_INSTANT may be assigned even if
instant ADD COLUMN was not committed. Changes to these page header fields
are not undo-logged, but changes to the 'default value record' are.
If the server is killed and restarted, the page header fields could
remain set even though no 'default value record' is present.
When the table becomes empty, the PAGE_INSTANT field and the
FIL_PAGE_TYPE can be reset and any 'default value record' be removed. */
#define PAGE_INSTANT 12
/** last insert direction: PAGE_LEFT, ....
In ROW_FORMAT=REDUNDANT tables created before MySQL 4.1.1 or MySQL 4.0.14,
this byte can be garbage. */
#define PAGE_DIRECTION_B 13
#define PAGE_N_DIRECTION 14 /* number of consecutive inserts to the same
direction */
#define PAGE_N_RECS 16 /* number of user records on the page */
MDEV-6076 Persistent AUTO_INCREMENT for InnoDB This should be functionally equivalent to WL#6204 in MySQL 8.0.0, with the notable difference that the file format changes are limited to repurposing a previously unused data field in B-tree pages. For persistent InnoDB tables, write the last used AUTO_INCREMENT value to the root page of the clustered index, in the previously unused (0) PAGE_MAX_TRX_ID field, now aliased as PAGE_ROOT_AUTO_INC. Unlike some other previously unused InnoDB data fields, this one was actually always zero-initialized, at least since MySQL 3.23.49. The writes to PAGE_ROOT_AUTO_INC are protected by SX or X latch on the root page. The SX latch will allow concurrent read access to the root page. (The field PAGE_ROOT_AUTO_INC will only be read on the first-time call to ha_innobase::open() from the SQL layer. The PAGE_ROOT_AUTO_INC can only be updated when executing SQL, so read/write races are not possible.) During INSERT, the PAGE_ROOT_AUTO_INC is updated by the low-level function btr_cur_search_to_nth_level(), adding no extra page access. [Adaptive hash index lookup will be disabled during INSERT.] If some rare UPDATE modifies an AUTO_INCREMENT column, the PAGE_ROOT_AUTO_INC will be adjusted in a separate mini-transaction in ha_innobase::update_row(). When a page is reorganized, we have to preserve the PAGE_ROOT_AUTO_INC field. During ALTER TABLE, the initial AUTO_INCREMENT value will be copied from the table. ALGORITHM=COPY and online log apply in LOCK=NONE will update PAGE_ROOT_AUTO_INC in real time. innodb_col_no(): Determine the dict_table_t::cols[] element index corresponding to a Field of a non-virtual column. (The MySQL 5.7 implementation of virtual columns breaks the 1:1 relationship between Field::field_index and dict_table_t::cols[]. Virtual columns are omitted from dict_table_t::cols[]. Therefore, we must translate the field_index of AUTO_INCREMENT columns into an index of dict_table_t::cols[].) Upgrade from old data files: By default, the AUTO_INCREMENT sequence in old data files would appear to be reset, because PAGE_MAX_TRX_ID or PAGE_ROOT_AUTO_INC would contain the value 0 in each clustered index page. In new data files, PAGE_ROOT_AUTO_INC can only be 0 if the table is empty or does not contain any AUTO_INCREMENT column. For backward compatibility, we use the old method of SELECT MAX(auto_increment_column) for initializing the sequence. btr_read_autoinc(): Read the AUTO_INCREMENT sequence from a new-format data file. btr_read_autoinc_with_fallback(): A variant of btr_read_autoinc() that will resort to reading MAX(auto_increment_column) for data files that did not use AUTO_INCREMENT yet. It was manually tested that during the execution of innodb.autoinc_persist the compatibility logic is not activated (for new files, PAGE_ROOT_AUTO_INC is never 0 in nonempty clustered index root pages). initialize_auto_increment(): Replaces ha_innobase::innobase_initialize_autoinc(). This initializes the AUTO_INCREMENT metadata. Only called from ha_innobase::open(). ha_innobase::info_low(): Do not try to lazily initialize dict_table_t::autoinc. It must already have been initialized by ha_innobase::open() or ha_innobase::create(). Note: The adjustments to class ha_innopart were not tested, because the source code (native InnoDB partitioning) is not being compiled.
2016-12-14 19:56:39 +02:00
/** The largest DB_TRX_ID that may have modified a record on the page;
Defined only in secondary index leaf pages and in change buffer leaf pages.
Otherwise written as 0. @see PAGE_ROOT_AUTO_INC */
#define PAGE_MAX_TRX_ID 18
/** The AUTO_INCREMENT value (on persistent clustered index root pages). */
#define PAGE_ROOT_AUTO_INC PAGE_MAX_TRX_ID
#define PAGE_HEADER_PRIV_END 26 /* end of private data structure of the page
header which are set in a page create */
/*----*/
#define PAGE_LEVEL 26 /* level of the node in an index tree; the
leaf level is the level 0. This field should
not be written to after page creation. */
#define PAGE_INDEX_ID 28 /* index id where the page belongs.
This field should not be written to after
page creation. */
#define PAGE_BTR_SEG_LEAF 36 /* file segment header for the leaf pages in
a B-tree: defined only on the root page of a
B-tree, but not in the root of an ibuf tree */
#define PAGE_BTR_IBUF_FREE_LIST PAGE_BTR_SEG_LEAF
#define PAGE_BTR_IBUF_FREE_LIST_NODE PAGE_BTR_SEG_LEAF
/* in the place of PAGE_BTR_SEG_LEAF and _TOP
there is a free list base node if the page is
the root page of an ibuf tree, and at the same
place is the free list node if the page is in
a free list */
#define PAGE_BTR_SEG_TOP (36 + FSEG_HEADER_SIZE)
/* file segment header for the non-leaf pages
in a B-tree: defined only on the root page of
a B-tree, but not in the root of an ibuf
tree */
/*----*/
#define PAGE_DATA (PAGE_HEADER + 36 + 2 * FSEG_HEADER_SIZE)
/* start of data on the page */
#define PAGE_OLD_INFIMUM (PAGE_DATA + 1 + REC_N_OLD_EXTRA_BYTES)
/* offset of the page infimum record on an
old-style page */
#define PAGE_OLD_SUPREMUM (PAGE_DATA + 2 + 2 * REC_N_OLD_EXTRA_BYTES + 8)
/* offset of the page supremum record on an
old-style page */
#define PAGE_OLD_SUPREMUM_END (PAGE_OLD_SUPREMUM + 9)
/* offset of the page supremum record end on
an old-style page */
#define PAGE_NEW_INFIMUM (PAGE_DATA + REC_N_NEW_EXTRA_BYTES)
/* offset of the page infimum record on a
new-style compact page */
#define PAGE_NEW_SUPREMUM (PAGE_DATA + 2 * REC_N_NEW_EXTRA_BYTES + 8)
/* offset of the page supremum record on a
new-style compact page */
#define PAGE_NEW_SUPREMUM_END (PAGE_NEW_SUPREMUM + 8)
/* offset of the page supremum record end on
a new-style compact page */
/*-----------------------------*/
/* Heap numbers */
#define PAGE_HEAP_NO_INFIMUM 0 /* page infimum */
#define PAGE_HEAP_NO_SUPREMUM 1 /* page supremum */
#define PAGE_HEAP_NO_USER_LOW 2 /* first user record in
creation (insertion) order,
not necessarily collation order;
this record may have been deleted */
/* Directions of cursor movement */
#define PAGE_LEFT 1
#define PAGE_RIGHT 2
#define PAGE_SAME_REC 3
#define PAGE_SAME_PAGE 4
#define PAGE_NO_DIRECTION 5
#ifndef UNIV_INNOCHECKSUM
/* PAGE DIRECTORY
==============
*/
typedef byte page_dir_slot_t;
typedef page_dir_slot_t page_dir_t;
/* Offset of the directory start down from the page end. We call the
slot with the highest file address directory start, as it points to
the first record in the list of records. */
#define PAGE_DIR FIL_PAGE_DATA_END
/* We define a slot in the page directory as two bytes */
#define PAGE_DIR_SLOT_SIZE 2
/* The offset of the physically lower end of the directory, counted from
page end, when the page is empty */
#define PAGE_EMPTY_DIR_START (PAGE_DIR + 2 * PAGE_DIR_SLOT_SIZE)
/* The maximum and minimum number of records owned by a directory slot. The
number may drop below the minimum in the first and the last slot in the
directory. */
#define PAGE_DIR_SLOT_MAX_N_OWNED 8
#define PAGE_DIR_SLOT_MIN_N_OWNED 4
2014-12-22 16:53:17 +02:00
extern my_bool srv_immediate_scrub_data_uncompressed;
#endif /* UNIV_INNOCHECKSUM */
2014-12-22 16:53:17 +02:00
/** Get the start of a page frame.
@param[in] ptr pointer within a page frame
@return start of the page frame */
MY_ATTRIBUTE((const))
inline
page_t*
page_align(const void* ptr)
{
return(static_cast<page_t*>(ut_align_down(ptr, UNIV_PAGE_SIZE)));
}
/** Gets the byte offset within a page frame.
@param[in] ptr pointer within a page frame
@return offset from the start of the page */
MY_ATTRIBUTE((const))
inline
ulint
page_offset(const void* ptr)
{
return(ut_align_offset(ptr, UNIV_PAGE_SIZE));
}
/** Determine whether an index page is not in ROW_FORMAT=REDUNDANT.
@param[in] page index page
@return nonzero if ROW_FORMAT is one of COMPACT,DYNAMIC,COMPRESSED
@retval 0 if ROW_FORMAT=REDUNDANT */
inline
byte
page_is_comp(const page_t* page)
{
ut_ad(!ut_align_offset(page, UNIV_ZIP_SIZE_MIN));
return(page[PAGE_HEADER + PAGE_N_HEAP] & 0x80);
}
/** Determine whether an index page is empty.
@param[in] page index page
@return whether the page is empty (PAGE_N_RECS = 0) */
inline
bool
page_is_empty(const page_t* page)
{
ut_ad(!ut_align_offset(page, UNIV_ZIP_SIZE_MIN));
return !*reinterpret_cast<const uint16_t*>(PAGE_HEADER + PAGE_N_RECS
+ page);
}
/** Determine whether an index page contains garbage.
@param[in] page index page
@return whether the page contains garbage (PAGE_GARBAGE is not 0) */
inline
bool
page_has_garbage(const page_t* page)
{
ut_ad(!ut_align_offset(page, UNIV_ZIP_SIZE_MIN));
return *reinterpret_cast<const uint16_t*>(PAGE_HEADER + PAGE_GARBAGE
+ page);
}
/** Determine whether an B-tree or R-tree index page is a leaf page.
@param[in] page index page
@return true if the page is a leaf (PAGE_LEVEL = 0) */
inline
bool
page_is_leaf(const page_t* page)
{
ut_ad(!ut_align_offset(page, UNIV_ZIP_SIZE_MIN));
return !*reinterpret_cast<const uint16_t*>(PAGE_HEADER + PAGE_LEVEL
+ page);
}
#ifndef UNIV_INNOCHECKSUM
/** Determine whether an index page record is not in ROW_FORMAT=REDUNDANT.
@param[in] rec record in an index page frame (not a copy)
@return nonzero if ROW_FORMAT is one of COMPACT,DYNAMIC,COMPRESSED
@retval 0 if ROW_FORMAT=REDUNDANT */
inline
byte
page_rec_is_comp(const byte* rec)
{
return(page_is_comp(page_align(rec)));
}
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
# ifdef UNIV_DEBUG
/** Determine if the record is the 'default row' pseudo-record
in the clustered index.
@param[in] rec leaf page record on an index page
@return whether the record is the 'default row' pseudo-record */
inline
bool
page_rec_is_default_row(const rec_t* rec)
{
return rec_get_info_bits(rec, page_rec_is_comp(rec))
& REC_INFO_MIN_REC_FLAG;
}
# endif /* UNIV_DEBUG */
/** Determine the offset of the infimum record on the page.
@param[in] page index page
@return offset of the infimum record in record list, relative from page */
inline
unsigned
page_get_infimum_offset(const page_t* page)
{
ut_ad(!page_offset(page));
return page_is_comp(page) ? PAGE_NEW_INFIMUM : PAGE_OLD_INFIMUM;
}
/** Determine the offset of the supremum record on the page.
@param[in] page index page
@return offset of the supremum record in record list, relative from page */
inline
unsigned
page_get_supremum_offset(const page_t* page)
{
ut_ad(!page_offset(page));
return page_is_comp(page) ? PAGE_NEW_SUPREMUM : PAGE_OLD_SUPREMUM;
}
/** Determine whether an index page record is a user record.
@param[in] offset record offset in the page
@retval true if a user record
@retval false if the infimum or supremum pseudo-record */
inline
bool
page_rec_is_user_rec_low(ulint offset)
{
compile_time_assert(PAGE_OLD_INFIMUM >= PAGE_NEW_INFIMUM);
compile_time_assert(PAGE_OLD_SUPREMUM >= PAGE_NEW_SUPREMUM);
compile_time_assert(PAGE_NEW_INFIMUM < PAGE_OLD_SUPREMUM);
compile_time_assert(PAGE_OLD_INFIMUM < PAGE_NEW_SUPREMUM);
compile_time_assert(PAGE_NEW_SUPREMUM < PAGE_OLD_SUPREMUM_END);
compile_time_assert(PAGE_OLD_SUPREMUM < PAGE_NEW_SUPREMUM_END);
ut_ad(offset >= PAGE_NEW_INFIMUM);
ut_ad(offset <= UNIV_PAGE_SIZE - PAGE_EMPTY_DIR_START);
return(offset != PAGE_NEW_SUPREMUM
&& offset != PAGE_NEW_INFIMUM
&& offset != PAGE_OLD_INFIMUM
&& offset != PAGE_OLD_SUPREMUM);
}
/** Determine if a record is the supremum record on an index page.
@param[in] offset record offset in an index page
@return true if the supremum record */
inline
bool
page_rec_is_supremum_low(ulint offset)
{
ut_ad(offset >= PAGE_NEW_INFIMUM);
ut_ad(offset <= UNIV_PAGE_SIZE - PAGE_EMPTY_DIR_START);
return(offset == PAGE_NEW_SUPREMUM || offset == PAGE_OLD_SUPREMUM);
}
/** Determine if a record is the infimum record on an index page.
@param[in] offset record offset in an index page
@return true if the infimum record */
inline
bool
page_rec_is_infimum_low(ulint offset)
{
ut_ad(offset >= PAGE_NEW_INFIMUM);
ut_ad(offset <= UNIV_PAGE_SIZE - PAGE_EMPTY_DIR_START);
return(offset == PAGE_NEW_INFIMUM || offset == PAGE_OLD_INFIMUM);
}
/** Determine whether an B-tree or R-tree index record is in a leaf page.
@param[in] rec index record in an index page
@return true if the record is in a leaf page */
inline
bool
page_rec_is_leaf(const page_t* rec)
{
const page_t* page = page_align(rec);
ut_ad(ulint(rec - page) >= page_get_infimum_offset(page));
bool leaf = page_is_leaf(page);
ut_ad(!page_rec_is_comp(rec)
|| !page_rec_is_user_rec_low(ulint(rec - page))
|| leaf == !rec_get_node_ptr_flag(rec));
return leaf;
}
/** Determine whether an index page record is a user record.
@param[in] rec record in an index page
@return true if a user record */
inline
bool
page_rec_is_user_rec(const rec_t* rec);
/** Determine whether an index page record is the supremum record.
@param[in] rec record in an index page
@return true if the supremum record */
inline
bool
page_rec_is_supremum(const rec_t* rec);
/** Determine whether an index page record is the infimum record.
@param[in] rec record in an index page
@return true if the infimum record */
inline
bool
page_rec_is_infimum(const rec_t* rec);
/*************************************************************//**
Returns the max trx id field value. */
UNIV_INLINE
trx_id_t
page_get_max_trx_id(
/*================*/
const page_t* page); /*!< in: page */
/*************************************************************//**
Sets the max trx id field value. */
void
page_set_max_trx_id(
/*================*/
buf_block_t* block, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page, or NULL */
trx_id_t trx_id, /*!< in: transaction id */
mtr_t* mtr); /*!< in/out: mini-transaction, or NULL */
/*************************************************************//**
Sets the max trx id field value if trx_id is bigger than the previous
value. */
UNIV_INLINE
void
page_update_max_trx_id(
/*===================*/
buf_block_t* block, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
trx_id_t trx_id, /*!< in: transaction id */
mtr_t* mtr); /*!< in/out: mini-transaction */
MDEV-6076 Persistent AUTO_INCREMENT for InnoDB This should be functionally equivalent to WL#6204 in MySQL 8.0.0, with the notable difference that the file format changes are limited to repurposing a previously unused data field in B-tree pages. For persistent InnoDB tables, write the last used AUTO_INCREMENT value to the root page of the clustered index, in the previously unused (0) PAGE_MAX_TRX_ID field, now aliased as PAGE_ROOT_AUTO_INC. Unlike some other previously unused InnoDB data fields, this one was actually always zero-initialized, at least since MySQL 3.23.49. The writes to PAGE_ROOT_AUTO_INC are protected by SX or X latch on the root page. The SX latch will allow concurrent read access to the root page. (The field PAGE_ROOT_AUTO_INC will only be read on the first-time call to ha_innobase::open() from the SQL layer. The PAGE_ROOT_AUTO_INC can only be updated when executing SQL, so read/write races are not possible.) During INSERT, the PAGE_ROOT_AUTO_INC is updated by the low-level function btr_cur_search_to_nth_level(), adding no extra page access. [Adaptive hash index lookup will be disabled during INSERT.] If some rare UPDATE modifies an AUTO_INCREMENT column, the PAGE_ROOT_AUTO_INC will be adjusted in a separate mini-transaction in ha_innobase::update_row(). When a page is reorganized, we have to preserve the PAGE_ROOT_AUTO_INC field. During ALTER TABLE, the initial AUTO_INCREMENT value will be copied from the table. ALGORITHM=COPY and online log apply in LOCK=NONE will update PAGE_ROOT_AUTO_INC in real time. innodb_col_no(): Determine the dict_table_t::cols[] element index corresponding to a Field of a non-virtual column. (The MySQL 5.7 implementation of virtual columns breaks the 1:1 relationship between Field::field_index and dict_table_t::cols[]. Virtual columns are omitted from dict_table_t::cols[]. Therefore, we must translate the field_index of AUTO_INCREMENT columns into an index of dict_table_t::cols[].) Upgrade from old data files: By default, the AUTO_INCREMENT sequence in old data files would appear to be reset, because PAGE_MAX_TRX_ID or PAGE_ROOT_AUTO_INC would contain the value 0 in each clustered index page. In new data files, PAGE_ROOT_AUTO_INC can only be 0 if the table is empty or does not contain any AUTO_INCREMENT column. For backward compatibility, we use the old method of SELECT MAX(auto_increment_column) for initializing the sequence. btr_read_autoinc(): Read the AUTO_INCREMENT sequence from a new-format data file. btr_read_autoinc_with_fallback(): A variant of btr_read_autoinc() that will resort to reading MAX(auto_increment_column) for data files that did not use AUTO_INCREMENT yet. It was manually tested that during the execution of innodb.autoinc_persist the compatibility logic is not activated (for new files, PAGE_ROOT_AUTO_INC is never 0 in nonempty clustered index root pages). initialize_auto_increment(): Replaces ha_innobase::innobase_initialize_autoinc(). This initializes the AUTO_INCREMENT metadata. Only called from ha_innobase::open(). ha_innobase::info_low(): Do not try to lazily initialize dict_table_t::autoinc. It must already have been initialized by ha_innobase::open() or ha_innobase::create(). Note: The adjustments to class ha_innopart were not tested, because the source code (native InnoDB partitioning) is not being compiled.
2016-12-14 19:56:39 +02:00
/** Persist the AUTO_INCREMENT value on a clustered index root page.
@param[in,out] block clustered index root page
@param[in] index clustered index
@param[in] autoinc next available AUTO_INCREMENT value
@param[in,out] mtr mini-transaction
@param[in] reset whether to reset the AUTO_INCREMENT
to a possibly smaller value than currently
exists in the page */
void
page_set_autoinc(
buf_block_t* block,
const dict_index_t* index MY_ATTRIBUTE((unused)),
ib_uint64_t autoinc,
mtr_t* mtr,
bool reset)
MY_ATTRIBUTE((nonnull));
/** Read the AUTO_INCREMENT value from a clustered index root page.
@param[in] page clustered index root page
@return the persisted AUTO_INCREMENT value */
MY_ATTRIBUTE((nonnull, warn_unused_result))
UNIV_INLINE
ib_uint64_t
page_get_autoinc(const page_t* page);
/*************************************************************//**
Returns the RTREE SPLIT SEQUENCE NUMBER (FIL_RTREE_SPLIT_SEQ_NUM).
@return SPLIT SEQUENCE NUMBER */
UNIV_INLINE
node_seq_t
page_get_ssn_id(
/*============*/
const page_t* page); /*!< in: page */
/*************************************************************//**
Sets the RTREE SPLIT SEQUENCE NUMBER field value */
UNIV_INLINE
void
page_set_ssn_id(
/*============*/
buf_block_t* block, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
node_seq_t ssn_id, /*!< in: split sequence id */
mtr_t* mtr); /*!< in/out: mini-transaction */
#endif /* !UNIV_INNOCHECKSUM */
/*************************************************************//**
Reads the given header field. */
UNIV_INLINE
2017-09-13 09:27:15 +03:00
uint16_t
page_header_get_field(
/*==================*/
const page_t* page, /*!< in: page */
ulint field); /*!< in: PAGE_N_DIR_SLOTS, ... */
#ifndef UNIV_INNOCHECKSUM
/*************************************************************//**
Sets the given header field. */
UNIV_INLINE
void
page_header_set_field(
/*==================*/
page_t* page, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
ulint field, /*!< in: PAGE_N_DIR_SLOTS, ... */
ulint val); /*!< in: value */
/*************************************************************//**
Returns the offset stored in the given header field.
@return offset from the start of the page, or 0 */
UNIV_INLINE
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
uint16_t
page_header_get_offs(
/*=================*/
const page_t* page, /*!< in: page */
ulint field) /*!< in: PAGE_FREE, ... */
MY_ATTRIBUTE((warn_unused_result));
/*************************************************************//**
Returns the pointer stored in the given header field, or NULL. */
#define page_header_get_ptr(page, field) \
(page_header_get_offs(page, field) \
? page + page_header_get_offs(page, field) : NULL)
/*************************************************************//**
Sets the pointer stored in the given header field. */
UNIV_INLINE
void
page_header_set_ptr(
/*================*/
page_t* page, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
ulint field, /*!< in/out: PAGE_FREE, ... */
const byte* ptr); /*!< in: pointer or NULL*/
/*************************************************************//**
Resets the last insert info field in the page header. Writes to mlog
about this operation. */
UNIV_INLINE
void
page_header_reset_last_insert(
/*==========================*/
page_t* page, /*!< in: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
mtr_t* mtr); /*!< in: mtr */
#define page_get_infimum_rec(page) ((page) + page_get_infimum_offset(page))
#define page_get_supremum_rec(page) ((page) + page_get_supremum_offset(page))
/************************************************************//**
Returns the nth record of the record list.
This is the inverse function of page_rec_get_n_recs_before().
@return nth record */
const rec_t*
page_rec_get_nth_const(
/*===================*/
const page_t* page, /*!< in: page */
ulint nth) /*!< in: nth record */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull, warn_unused_result));
/************************************************************//**
Returns the nth record of the record list.
This is the inverse function of page_rec_get_n_recs_before().
@return nth record */
UNIV_INLINE
rec_t*
page_rec_get_nth(
/*=============*/
page_t* page, /*< in: page */
ulint nth) /*!< in: nth record */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull, warn_unused_result));
/************************************************************//**
Returns the middle record of the records on the page. If there is an
even number of records in the list, returns the first record of the
upper half-list.
@return middle record */
UNIV_INLINE
rec_t*
page_get_middle_rec(
/*================*/
page_t* page) /*!< in: page */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull, warn_unused_result));
/*************************************************************//**
Gets the page number.
@return page number */
UNIV_INLINE
ulint
page_get_page_no(
/*=============*/
const page_t* page); /*!< in: page */
/*************************************************************//**
Gets the tablespace identifier.
@return space id */
UNIV_INLINE
ulint
page_get_space_id(
/*==============*/
const page_t* page); /*!< in: page */
/*************************************************************//**
Gets the number of user records on page (the infimum and supremum records
are not user records).
@return number of user records */
UNIV_INLINE
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
uint16_t
page_get_n_recs(
/*============*/
const page_t* page); /*!< in: index page */
/***************************************************************//**
Returns the number of records before the given record in chain.
The number includes infimum and supremum records.
This is the inverse function of page_rec_get_nth().
@return number of records */
ulint
page_rec_get_n_recs_before(
/*=======================*/
const rec_t* rec); /*!< in: the physical record */
/*************************************************************//**
Gets the number of records in the heap.
@return number of user records */
UNIV_INLINE
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
uint16_t
page_dir_get_n_heap(
/*================*/
const page_t* page); /*!< in: index page */
/*************************************************************//**
Sets the number of records in the heap. */
UNIV_INLINE
void
page_dir_set_n_heap(
/*================*/
page_t* page, /*!< in/out: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL.
Note that the size of the dense page directory
in the compressed page trailer is
n_heap * PAGE_ZIP_DIR_SLOT_SIZE. */
ulint n_heap);/*!< in: number of records */
/*************************************************************//**
Gets the number of dir slots in directory.
@return number of slots */
UNIV_INLINE
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
uint16_t
page_dir_get_n_slots(
/*=================*/
const page_t* page); /*!< in: index page */
/*************************************************************//**
Sets the number of dir slots in directory. */
UNIV_INLINE
void
page_dir_set_n_slots(
/*=================*/
page_t* page, /*!< in/out: page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be updated, or NULL */
ulint n_slots);/*!< in: number of slots */
#ifdef UNIV_DEBUG
/*************************************************************//**
Gets pointer to nth directory slot.
@return pointer to dir slot */
UNIV_INLINE
page_dir_slot_t*
page_dir_get_nth_slot(
/*==================*/
const page_t* page, /*!< in: index page */
ulint n); /*!< in: position */
#else /* UNIV_DEBUG */
# define page_dir_get_nth_slot(page, n) \
((page) + (UNIV_PAGE_SIZE - PAGE_DIR \
- (n + 1) * PAGE_DIR_SLOT_SIZE))
#endif /* UNIV_DEBUG */
/**************************************************************//**
Used to check the consistency of a record on a page.
@return TRUE if succeed */
UNIV_INLINE
ibool
page_rec_check(
/*===========*/
const rec_t* rec); /*!< in: record */
/***************************************************************//**
Gets the record pointed to by a directory slot.
@return pointer to record */
UNIV_INLINE
const rec_t*
page_dir_slot_get_rec(
/*==================*/
const page_dir_slot_t* slot); /*!< in: directory slot */
/***************************************************************//**
This is used to set the record offset in a directory slot. */
UNIV_INLINE
void
page_dir_slot_set_rec(
/*==================*/
page_dir_slot_t* slot, /*!< in: directory slot */
rec_t* rec); /*!< in: record on the page */
/***************************************************************//**
Gets the number of records owned by a directory slot.
@return number of records */
UNIV_INLINE
ulint
page_dir_slot_get_n_owned(
/*======================*/
const page_dir_slot_t* slot); /*!< in: page directory slot */
/***************************************************************//**
This is used to set the owned records field of a directory slot. */
UNIV_INLINE
void
page_dir_slot_set_n_owned(
/*======================*/
page_dir_slot_t*slot, /*!< in/out: directory slot */
page_zip_des_t* page_zip,/*!< in/out: compressed page, or NULL */
ulint n); /*!< in: number of records owned by the slot */
/************************************************************//**
Calculates the space reserved for directory slots of a given
number of records. The exact value is a fraction number
n * PAGE_DIR_SLOT_SIZE / PAGE_DIR_SLOT_MIN_N_OWNED, and it is
rounded upwards to an integer. */
UNIV_INLINE
ulint
page_dir_calc_reserved_space(
/*=========================*/
ulint n_recs); /*!< in: number of records */
/***************************************************************//**
Looks for the directory slot which owns the given record.
@return the directory slot number */
ulint
page_dir_find_owner_slot(
/*=====================*/
const rec_t* rec); /*!< in: the physical record */
/***************************************************************//**
Returns the heap number of a record.
@return heap number */
UNIV_INLINE
ulint
page_rec_get_heap_no(
/*=================*/
const rec_t* rec); /*!< in: the physical record */
/** Determine whether a page is an index root page.
@param[in] page page frame
@return true if the page is a root page of an index */
UNIV_INLINE
bool
page_is_root(
const page_t* page)
MY_ATTRIBUTE((warn_unused_result));
/************************************************************//**
Gets the pointer to the next record on the page.
@return pointer to next record */
UNIV_INLINE
const rec_t*
page_rec_get_next_low(
/*==================*/
const rec_t* rec, /*!< in: pointer to record */
ulint comp); /*!< in: nonzero=compact page layout */
/************************************************************//**
Gets the pointer to the next record on the page.
@return pointer to next record */
UNIV_INLINE
rec_t*
page_rec_get_next(
/*==============*/
rec_t* rec); /*!< in: pointer to record */
/************************************************************//**
Gets the pointer to the next record on the page.
@return pointer to next record */
UNIV_INLINE
const rec_t*
page_rec_get_next_const(
/*====================*/
const rec_t* rec); /*!< in: pointer to record */
/************************************************************//**
Gets the pointer to the next non delete-marked record on the page.
If all subsequent records are delete-marked, then this function
will return the supremum record.
@return pointer to next non delete-marked record or pointer to supremum */
UNIV_INLINE
const rec_t*
page_rec_get_next_non_del_marked(
/*=============================*/
const rec_t* rec); /*!< in: pointer to record */
/************************************************************//**
Sets the pointer to the next record on the page. */
UNIV_INLINE
void
page_rec_set_next(
/*==============*/
rec_t* rec, /*!< in: pointer to record,
must not be page supremum */
const rec_t* next); /*!< in: pointer to next record,
must not be page infimum */
/************************************************************//**
Gets the pointer to the previous record.
@return pointer to previous record */
UNIV_INLINE
const rec_t*
page_rec_get_prev_const(
/*====================*/
const rec_t* rec); /*!< in: pointer to record, must not be page
infimum */
/************************************************************//**
Gets the pointer to the previous record.
@return pointer to previous record */
UNIV_INLINE
rec_t*
page_rec_get_prev(
/*==============*/
rec_t* rec); /*!< in: pointer to record,
must not be page infimum */
/************************************************************//**
true if the record is the first user record on a page.
@return true if the first user record */
UNIV_INLINE
bool
page_rec_is_first(
/*==============*/
const rec_t* rec, /*!< in: record */
const page_t* page) /*!< in: page */
MY_ATTRIBUTE((warn_unused_result));
/************************************************************//**
true if the record is the second user record on a page.
@return true if the second user record */
UNIV_INLINE
bool
page_rec_is_second(
/*===============*/
const rec_t* rec, /*!< in: record */
const page_t* page) /*!< in: page */
MY_ATTRIBUTE((warn_unused_result));
/************************************************************//**
true if the record is the last user record on a page.
@return true if the last user record */
UNIV_INLINE
bool
page_rec_is_last(
/*=============*/
const rec_t* rec, /*!< in: record */
const page_t* page) /*!< in: page */
MY_ATTRIBUTE((warn_unused_result));
/************************************************************//**
true if the record is the second last user record on a page.
@return true if the second last user record */
UNIV_INLINE
bool
page_rec_is_second_last(
/*====================*/
const rec_t* rec, /*!< in: record */
const page_t* page) /*!< in: page */
MY_ATTRIBUTE((warn_unused_result));
/***************************************************************//**
Looks for the record which owns the given record.
@return the owner record */
UNIV_INLINE
rec_t*
page_rec_find_owner_rec(
/*====================*/
rec_t* rec); /*!< in: the physical record */
/***********************************************************************//**
Write a 32-bit field in a data dictionary record. */
UNIV_INLINE
void
page_rec_write_field(
/*=================*/
rec_t* rec, /*!< in/out: record to update */
ulint i, /*!< in: index of the field to update */
ulint val, /*!< in: value to write */
mtr_t* mtr) /*!< in/out: mini-transaction */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull));
/************************************************************//**
Returns the maximum combined size of records which can be inserted on top
of record heap.
@return maximum combined size for inserted records */
UNIV_INLINE
ulint
page_get_max_insert_size(
/*=====================*/
const page_t* page, /*!< in: index page */
ulint n_recs);/*!< in: number of records */
/************************************************************//**
Returns the maximum combined size of records which can be inserted on top
of record heap if page is first reorganized.
@return maximum combined size for inserted records */
UNIV_INLINE
ulint
page_get_max_insert_size_after_reorganize(
/*======================================*/
const page_t* page, /*!< in: index page */
ulint n_recs);/*!< in: number of records */
/*************************************************************//**
Calculates free space if a page is emptied.
@return free space */
UNIV_INLINE
ulint
page_get_free_space_of_empty(
/*=========================*/
ulint comp) /*!< in: nonzero=compact page format */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((const));
/**********************************************************//**
Returns the base extra size of a physical record. This is the
size of the fixed header, independent of the record size.
@return REC_N_NEW_EXTRA_BYTES or REC_N_OLD_EXTRA_BYTES */
UNIV_INLINE
ulint
page_rec_get_base_extra_size(
/*=========================*/
const rec_t* rec); /*!< in: physical record */
/************************************************************//**
Returns the sum of the sizes of the records in the record list
excluding the infimum and supremum records.
@return data in bytes */
UNIV_INLINE
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
uint16_t
page_get_data_size(
/*===============*/
const page_t* page); /*!< in: index page */
/************************************************************//**
Allocates a block of memory from the head of the free list
of an index page. */
UNIV_INLINE
void
page_mem_alloc_free(
/*================*/
page_t* page, /*!< in/out: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page with enough
space available for inserting the record,
or NULL */
rec_t* next_rec,/*!< in: pointer to the new head of the
free record list */
ulint need); /*!< in: number of bytes allocated */
/************************************************************//**
Allocates a block of memory from the heap of an index page.
@return pointer to start of allocated buffer, or NULL if allocation fails */
byte*
page_mem_alloc_heap(
/*================*/
page_t* page, /*!< in/out: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page with enough
space available for inserting the record,
or NULL */
ulint need, /*!< in: total number of bytes needed */
ulint* heap_no);/*!< out: this contains the heap number
of the allocated record
if allocation succeeds */
/************************************************************//**
Puts a record to free list. */
UNIV_INLINE
void
page_mem_free(
/*==========*/
page_t* page, /*!< in/out: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page,
or NULL */
rec_t* rec, /*!< in: pointer to the (origin of)
record */
const dict_index_t* index, /*!< in: index of rec */
const ulint* offsets);/*!< in: array returned by
rec_get_offsets() */
MDEV-11369 Instant ADD COLUMN for InnoDB For InnoDB tables, adding, dropping and reordering columns has required a rebuild of the table and all its indexes. Since MySQL 5.6 (and MariaDB 10.0) this has been supported online (LOCK=NONE), allowing concurrent modification of the tables. This work revises the InnoDB ROW_FORMAT=REDUNDANT, ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC so that columns can be appended instantaneously, with only minor changes performed to the table structure. The counter innodb_instant_alter_column in INFORMATION_SCHEMA.GLOBAL_STATUS is incremented whenever a table rebuild operation is converted into an instant ADD COLUMN operation. ROW_FORMAT=COMPRESSED tables will not support instant ADD COLUMN. Some usability limitations will be addressed in subsequent work: MDEV-13134 Introduce ALTER TABLE attributes ALGORITHM=NOCOPY and ALGORITHM=INSTANT MDEV-14016 Allow instant ADD COLUMN, ADD INDEX, LOCK=NONE The format of the clustered index (PRIMARY KEY) is changed as follows: (1) The FIL_PAGE_TYPE of the root page will be FIL_PAGE_TYPE_INSTANT, and a new field PAGE_INSTANT will contain the original number of fields in the clustered index ('core' fields). If instant ADD COLUMN has not been used or the table becomes empty, or the very first instant ADD COLUMN operation is rolled back, the fields PAGE_INSTANT and FIL_PAGE_TYPE will be reset to 0 and FIL_PAGE_INDEX. (2) A special 'default row' record is inserted into the leftmost leaf, between the page infimum and the first user record. This record is distinguished by the REC_INFO_MIN_REC_FLAG, and it is otherwise in the same format as records that contain values for the instantly added columns. This 'default row' always has the same number of fields as the clustered index according to the table definition. The values of 'core' fields are to be ignored. For other fields, the 'default row' will contain the default values as they were during the ALTER TABLE statement. (If the column default values are changed later, those values will only be stored in the .frm file. The 'default row' will contain the original evaluated values, which must be the same for every row.) The 'default row' must be completely hidden from higher-level access routines. Assertions have been added to ensure that no 'default row' is ever present in the adaptive hash index or in locked records. The 'default row' is never delete-marked. (3) In clustered index leaf page records, the number of fields must reside between the number of 'core' fields (dict_index_t::n_core_fields introduced in this work) and dict_index_t::n_fields. If the number of fields is less than dict_index_t::n_fields, the missing fields are replaced with the column value of the 'default row'. Note: The number of fields in the record may shrink if some of the last instantly added columns are updated to the value that is in the 'default row'. The function btr_cur_trim() implements this 'compression' on update and rollback; dtuple::trim() implements it on insert. (4) In ROW_FORMAT=COMPACT and ROW_FORMAT=DYNAMIC records, the new status value REC_STATUS_COLUMNS_ADDED will indicate the presence of a new record header that will encode n_fields-n_core_fields-1 in 1 or 2 bytes. (In ROW_FORMAT=REDUNDANT records, the record header always explicitly encodes the number of fields.) We introduce the undo log record type TRX_UNDO_INSERT_DEFAULT for covering the insert of the 'default row' record when instant ADD COLUMN is used for the first time. Subsequent instant ADD COLUMN can use TRX_UNDO_UPD_EXIST_REC. This is joint work with Vin Chen (陈福荣) from Tencent. The design that was discussed in April 2017 would not have allowed import or export of data files, because instead of the 'default row' it would have introduced a data dictionary table. The test rpl.rpl_alter_instant is exactly as contributed in pull request #408. The test innodb.instant_alter is based on a contributed test. The redo log record format changes for ROW_FORMAT=DYNAMIC and ROW_FORMAT=COMPACT are as contributed. (With this change present, crash recovery from MariaDB 10.3.1 will fail in spectacular ways!) Also the semantics of higher-level redo log records that modify the PAGE_INSTANT field is changed. The redo log format version identifier was already changed to LOG_HEADER_FORMAT_CURRENT=103 in MariaDB 10.3.1. Everything else has been rewritten by me. Thanks to Elena Stepanova, the code has been tested extensively. When rolling back an instant ADD COLUMN operation, we must empty the PAGE_FREE list after deleting or shortening the 'default row' record, by calling either btr_page_empty() or btr_page_reorganize(). We must know the size of each entry in the PAGE_FREE list. If rollback left a freed copy of the 'default row' in the PAGE_FREE list, we would be unable to determine its size (if it is in ROW_FORMAT=COMPACT or ROW_FORMAT=DYNAMIC) because it would contain more fields than the rolled-back definition of the clustered index. UNIV_SQL_DEFAULT: A new special constant that designates an instantly added column that is not present in the clustered index record. len_is_stored(): Check if a length is an actual length. There are two magic length values: UNIV_SQL_DEFAULT, UNIV_SQL_NULL. dict_col_t::def_val: The 'default row' value of the column. If the column is not added instantly, def_val.len will be UNIV_SQL_DEFAULT. dict_col_t: Add the accessors is_virtual(), is_nullable(), is_instant(), instant_value(). dict_col_t::remove_instant(): Remove the 'instant ADD' status of a column. dict_col_t::name(const dict_table_t& table): Replaces dict_table_get_col_name(). dict_index_t::n_core_fields: The original number of fields. For secondary indexes and if instant ADD COLUMN has not been used, this will be equal to dict_index_t::n_fields. dict_index_t::n_core_null_bytes: Number of bytes needed to represent the null flags; usually equal to UT_BITS_IN_BYTES(n_nullable). dict_index_t::NO_CORE_NULL_BYTES: Magic value signalling that n_core_null_bytes was not initialized yet from the clustered index root page. dict_index_t: Add the accessors is_instant(), is_clust(), get_n_nullable(), instant_field_value(). dict_index_t::instant_add_field(): Adjust clustered index metadata for instant ADD COLUMN. dict_index_t::remove_instant(): Remove the 'instant ADD' status of a clustered index when the table becomes empty, or the very first instant ADD COLUMN operation is rolled back. dict_table_t: Add the accessors is_instant(), is_temporary(), supports_instant(). dict_table_t::instant_add_column(): Adjust metadata for instant ADD COLUMN. dict_table_t::rollback_instant(): Adjust metadata on the rollback of instant ADD COLUMN. prepare_inplace_alter_table_dict(): First create the ctx->new_table, and only then decide if the table really needs to be rebuilt. We must split the creation of table or index metadata from the creation of the dictionary table records and the creation of the data. In this way, we can transform a table-rebuilding operation into an instant ADD COLUMN operation. Dictionary objects will only be added to cache when table rebuilding or index creation is needed. The ctx->instant_table will never be added to cache. dict_table_t::add_to_cache(): Modified and renamed from dict_table_add_to_cache(). Do not modify the table metadata. Let the callers invoke dict_table_add_system_columns() and if needed, set can_be_evicted. dict_create_sys_tables_tuple(), dict_create_table_step(): Omit the system columns (which will now exist in the dict_table_t object already at this point). dict_create_table_step(): Expect the callers to invoke dict_table_add_system_columns(). pars_create_table(): Before creating the table creation execution graph, invoke dict_table_add_system_columns(). row_create_table_for_mysql(): Expect all callers to invoke dict_table_add_system_columns(). create_index_dict(): Replaces row_merge_create_index_graph(). innodb_update_n_cols(): Renamed from innobase_update_n_virtual(). Call my_error() if an error occurs. btr_cur_instant_init(), btr_cur_instant_init_low(), btr_cur_instant_root_init(): Load additional metadata from the clustered index and set dict_index_t::n_core_null_bytes. This is invoked when table metadata is first loaded into the data dictionary. dict_boot(): Initialize n_core_null_bytes for the four hard-coded dictionary tables. dict_create_index_step(): Initialize n_core_null_bytes. This is executed as part of CREATE TABLE. dict_index_build_internal_clust(): Initialize n_core_null_bytes to NO_CORE_NULL_BYTES if table->supports_instant(). row_create_index_for_mysql(): Initialize n_core_null_bytes for CREATE TEMPORARY TABLE. commit_cache_norebuild(): Call the code to rename or enlarge columns in the cache only if instant ADD COLUMN is not being used. (Instant ADD COLUMN would copy all column metadata from instant_table to old_table, including the names and lengths.) PAGE_INSTANT: A new 13-bit field for storing dict_index_t::n_core_fields. This is repurposing the 16-bit field PAGE_DIRECTION, of which only the least significant 3 bits were used. The original byte containing PAGE_DIRECTION will be accessible via the new constant PAGE_DIRECTION_B. page_get_instant(), page_set_instant(): Accessors for the PAGE_INSTANT. page_ptr_get_direction(), page_get_direction(), page_ptr_set_direction(): Accessors for PAGE_DIRECTION. page_direction_reset(): Reset PAGE_DIRECTION, PAGE_N_DIRECTION. page_direction_increment(): Increment PAGE_N_DIRECTION and set PAGE_DIRECTION. rec_get_offsets(): Use the 'leaf' parameter for non-debug purposes, and assume that heap_no is always set. Initialize all dict_index_t::n_fields for ROW_FORMAT=REDUNDANT records, even if the record contains fewer fields. rec_offs_make_valid(): Add the parameter 'leaf'. rec_copy_prefix_to_dtuple(): Assert that the tuple is only built on the core fields. Instant ADD COLUMN only applies to the clustered index, and we should never build a search key that has more than the PRIMARY KEY and possibly DB_TRX_ID,DB_ROLL_PTR. All these columns are always present. dict_index_build_data_tuple(): Remove assertions that would be duplicated in rec_copy_prefix_to_dtuple(). rec_init_offsets(): Support ROW_FORMAT=REDUNDANT records whose number of fields is between n_core_fields and n_fields. cmp_rec_rec_with_match(): Implement the comparison between two MIN_REC_FLAG records. trx_t::in_rollback: Make the field available in non-debug builds. trx_start_for_ddl_low(): Remove dangerous error-tolerance. A dictionary transaction must be flagged as such before it has generated any undo log records. This is because trx_undo_assign_undo() will mark the transaction as a dictionary transaction in the undo log header right before the very first undo log record is being written. btr_index_rec_validate(): Account for instant ADD COLUMN row_undo_ins_remove_clust_rec(): On the rollback of an insert into SYS_COLUMNS, revert instant ADD COLUMN in the cache by removing the last column from the table and the clustered index. row_search_on_row_ref(), row_undo_mod_parse_undo_rec(), row_undo_mod(), trx_undo_update_rec_get_update(): Handle the 'default row' as a special case. dtuple_t::trim(index): Omit a redundant suffix of an index tuple right before insert or update. After instant ADD COLUMN, if the last fields of a clustered index tuple match the 'default row', there is no need to store them. While trimming the entry, we must hold a page latch, so that the table cannot be emptied and the 'default row' be deleted. btr_cur_optimistic_update(), btr_cur_pessimistic_update(), row_upd_clust_rec_by_insert(), row_ins_clust_index_entry_low(): Invoke dtuple_t::trim() if needed. row_ins_clust_index_entry(): Restore dtuple_t::n_fields after calling row_ins_clust_index_entry_low(). rec_get_converted_size(), rec_get_converted_size_comp(): Allow the number of fields to be between n_core_fields and n_fields. Do not support infimum,supremum. They are never supposed to be stored in dtuple_t, because page creation nowadays uses a lower-level method for initializing them. rec_convert_dtuple_to_rec_comp(): Assign the status bits based on the number of fields. btr_cur_trim(): In an update, trim the index entry as needed. For the 'default row', handle rollback specially. For user records, omit fields that match the 'default row'. btr_cur_optimistic_delete_func(), btr_cur_pessimistic_delete(): Skip locking and adaptive hash index for the 'default row'. row_log_table_apply_convert_mrec(): Replace 'default row' values if needed. In the temporary file that is applied by row_log_table_apply(), we must identify whether the records contain the extra header for instantly added columns. For now, we will allocate an additional byte for this for ROW_T_INSERT and ROW_T_UPDATE records when the source table has been subject to instant ADD COLUMN. The ROW_T_DELETE records are fine, as they will be converted and will only contain 'core' columns (PRIMARY KEY and some system columns) that are converted from dtuple_t. rec_get_converted_size_temp(), rec_init_offsets_temp(), rec_convert_dtuple_to_temp(): Add the parameter 'status'. REC_INFO_DEFAULT_ROW = REC_INFO_MIN_REC_FLAG | REC_STATUS_COLUMNS_ADDED: An info_bits constant for distinguishing the 'default row' record. rec_comp_status_t: An enum of the status bit values. rec_leaf_format: An enum that replaces the bool parameter of rec_init_offsets_comp_ordinary().
2017-10-06 07:00:05 +03:00
/** Read the PAGE_DIRECTION field from a byte.
@param[in] ptr pointer to PAGE_DIRECTION_B
@return the value of the PAGE_DIRECTION field */
inline
byte
page_ptr_get_direction(const byte* ptr);
/** Set the PAGE_DIRECTION field.
@param[in] ptr pointer to PAGE_DIRECTION_B
@param[in] dir the value of the PAGE_DIRECTION field */
inline
void
page_ptr_set_direction(byte* ptr, byte dir);
/** Read the PAGE_DIRECTION field.
@param[in] page index page
@return the value of the PAGE_DIRECTION field */
inline
byte
page_get_direction(const page_t* page)
{
return page_ptr_get_direction(PAGE_HEADER + PAGE_DIRECTION_B + page);
}
/** Read the PAGE_INSTANT field.
@param[in] page index page
@return the value of the PAGE_INSTANT field */
inline
uint16_t
page_get_instant(const page_t* page);
/** Assign the PAGE_INSTANT field.
@param[in,out] page clustered index root page
@param[in] n original number of clustered index fields
@param[in,out] mtr mini-transaction */
inline
void
page_set_instant(page_t* page, unsigned n, mtr_t* mtr);
/**********************************************************//**
Create an uncompressed B-tree index page.
@return pointer to the page */
page_t*
page_create(
/*========*/
buf_block_t* block, /*!< in: a buffer block where the
page is created */
mtr_t* mtr, /*!< in: mini-transaction handle */
ulint comp, /*!< in: nonzero=compact page format */
bool is_rtree); /*!< in: if creating R-tree page */
/**********************************************************//**
Create a compressed B-tree index page.
@return pointer to the page */
page_t*
page_create_zip(
/*============*/
buf_block_t* block, /*!< in/out: a buffer frame
where the page is created */
dict_index_t* index, /*!< in: the index of the
page, or NULL when applying
TRUNCATE log
record during recovery */
ulint level, /*!< in: the B-tree level of
the page */
trx_id_t max_trx_id, /*!< in: PAGE_MAX_TRX_ID */
const redo_page_compress_t* page_comp_info,
/*!< in: used for applying
TRUNCATE log
record during recovery */
mtr_t* mtr); /*!< in/out: mini-transaction
handle */
/**********************************************************//**
Empty a previously created B-tree index page. */
void
page_create_empty(
/*==============*/
buf_block_t* block, /*!< in/out: B-tree block */
dict_index_t* index, /*!< in: the index of the page */
mtr_t* mtr) /*!< in/out: mini-transaction */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull(1,2)));
/*************************************************************//**
Differs from page_copy_rec_list_end, because this function does not
touch the lock table and max trx id on page or compress the page.
IMPORTANT: The caller will have to update IBUF_BITMAP_FREE
if new_block is a compressed leaf page in a secondary index.
This has to be done either within the same mini-transaction,
or by invoking ibuf_reset_free_bits() before mtr_commit(). */
void
page_copy_rec_list_end_no_locks(
/*============================*/
buf_block_t* new_block, /*!< in: index page to copy to */
buf_block_t* block, /*!< in: index page of rec */
rec_t* rec, /*!< in: record on page */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr); /*!< in: mtr */
/*************************************************************//**
Copies records from page to new_page, from the given record onward,
including that record. Infimum and supremum records are not copied.
The records are copied to the start of the record list on new_page.
IMPORTANT: The caller will have to update IBUF_BITMAP_FREE
if new_block is a compressed leaf page in a secondary index.
This has to be done either within the same mini-transaction,
or by invoking ibuf_reset_free_bits() before mtr_commit().
@return pointer to the original successor of the infimum record on
new_page, or NULL on zip overflow (new_block will be decompressed) */
rec_t*
page_copy_rec_list_end(
/*===================*/
buf_block_t* new_block, /*!< in/out: index page to copy to */
buf_block_t* block, /*!< in: index page containing rec */
rec_t* rec, /*!< in: record on page */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Copies records from page to new_page, up to the given record, NOT
including that record. Infimum and supremum records are not copied.
The records are copied to the end of the record list on new_page.
IMPORTANT: The caller will have to update IBUF_BITMAP_FREE
if new_block is a compressed leaf page in a secondary index.
This has to be done either within the same mini-transaction,
or by invoking ibuf_reset_free_bits() before mtr_commit().
@return pointer to the original predecessor of the supremum record on
new_page, or NULL on zip overflow (new_block will be decompressed) */
rec_t*
page_copy_rec_list_start(
/*=====================*/
buf_block_t* new_block, /*!< in/out: index page to copy to */
buf_block_t* block, /*!< in: index page containing rec */
rec_t* rec, /*!< in: record on page */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Deletes records from a page from a given record onward, including that record.
The infimum and supremum records are not deleted. */
void
page_delete_rec_list_end(
/*=====================*/
rec_t* rec, /*!< in: pointer to record on page */
buf_block_t* block, /*!< in: buffer block of the page */
dict_index_t* index, /*!< in: record descriptor */
ulint n_recs, /*!< in: number of records to delete,
or ULINT_UNDEFINED if not known */
ulint size, /*!< in: the sum of the sizes of the
records in the end of the chain to
delete, or ULINT_UNDEFINED if not known */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Deletes records from page, up to the given record, NOT including
that record. Infimum and supremum records are not deleted. */
void
page_delete_rec_list_start(
/*=======================*/
rec_t* rec, /*!< in: record on page */
buf_block_t* block, /*!< in: buffer block of the page */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull));
/*************************************************************//**
Moves record list end to another page. Moved records include
split_rec.
IMPORTANT: The caller will have to update IBUF_BITMAP_FREE
if new_block is a compressed leaf page in a secondary index.
This has to be done either within the same mini-transaction,
or by invoking ibuf_reset_free_bits() before mtr_commit().
@return TRUE on success; FALSE on compression failure (new_block will
be decompressed) */
ibool
page_move_rec_list_end(
/*===================*/
buf_block_t* new_block, /*!< in/out: index page where to move */
buf_block_t* block, /*!< in: index page from where to move */
rec_t* split_rec, /*!< in: first record to move */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull(1, 2, 4, 5)));
/*************************************************************//**
Moves record list start to another page. Moved records do not include
split_rec.
IMPORTANT: The caller will have to update IBUF_BITMAP_FREE
if new_block is a compressed leaf page in a secondary index.
This has to be done either within the same mini-transaction,
or by invoking ibuf_reset_free_bits() before mtr_commit().
@return TRUE on success; FALSE on compression failure */
ibool
page_move_rec_list_start(
/*=====================*/
buf_block_t* new_block, /*!< in/out: index page where to move */
buf_block_t* block, /*!< in/out: page containing split_rec */
rec_t* split_rec, /*!< in: first record not to move */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr) /*!< in: mtr */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull(1, 2, 4, 5)));
/****************************************************************//**
Splits a directory slot which owns too many records. */
void
page_dir_split_slot(
/*================*/
page_t* page, /*!< in: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page whose
uncompressed part will be written, or NULL */
ulint slot_no)/*!< in: the directory slot */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull(1)));
/*************************************************************//**
Tries to balance the given directory slot with too few records
with the upper neighbor, so that there are at least the minimum number
of records owned by the slot; this may result in the merging of
two slots. */
void
page_dir_balance_slot(
/*==================*/
page_t* page, /*!< in/out: index page */
page_zip_des_t* page_zip,/*!< in/out: compressed page, or NULL */
ulint slot_no)/*!< in: the directory slot */
2016-06-21 14:21:03 +02:00
MY_ATTRIBUTE((nonnull(1)));
/**********************************************************//**
Parses a log record of a record list end or start deletion.
@return end of log record or NULL */
byte*
page_parse_delete_rec_list(
/*=======================*/
mlog_id_t type, /*!< in: MLOG_LIST_END_DELETE,
MLOG_LIST_START_DELETE,
MLOG_COMP_LIST_END_DELETE or
MLOG_COMP_LIST_START_DELETE */
byte* ptr, /*!< in: buffer */
byte* end_ptr,/*!< in: buffer end */
buf_block_t* block, /*!< in/out: buffer block or NULL */
dict_index_t* index, /*!< in: record descriptor */
mtr_t* mtr); /*!< in: mtr or NULL */
/** Parses a redo log record of creating a page.
@param[in,out] block buffer block, or NULL
@param[in] comp nonzero=compact page format
@param[in] is_rtree whether it is rtree page */
void
page_parse_create(
buf_block_t* block,
ulint comp,
bool is_rtree);
/************************************************************//**
Prints record contents including the data relevant only in
the index page context. */
void
page_rec_print(
/*===========*/
const rec_t* rec, /*!< in: physical record */
const ulint* offsets);/*!< in: record descriptor */
# ifdef UNIV_BTR_PRINT
/***************************************************************//**
This is used to print the contents of the directory for
debugging purposes. */
void
page_dir_print(
/*===========*/
page_t* page, /*!< in: index page */
ulint pr_n); /*!< in: print n first and n last entries */
/***************************************************************//**
This is used to print the contents of the page record list for
debugging purposes. */
void
page_print_list(
/*============*/
buf_block_t* block, /*!< in: index page */
dict_index_t* index, /*!< in: dictionary index of the page */
ulint pr_n); /*!< in: print n first and n last entries */
/***************************************************************//**
Prints the info in a page header. */
void
page_header_print(
/*==============*/
const page_t* page); /*!< in: index page */
/***************************************************************//**
This is used to print the contents of the page for
debugging purposes. */
void
page_print(
/*=======*/
buf_block_t* block, /*!< in: index page */
dict_index_t* index, /*!< in: dictionary index of the page */
ulint dn, /*!< in: print dn first and last entries
in directory */
ulint rn); /*!< in: print rn first and last records
in directory */
# endif /* UNIV_BTR_PRINT */
/***************************************************************//**
The following is used to validate a record on a page. This function
differs from rec_validate as it can also check the n_owned field and
the heap_no field.
@return TRUE if ok */
ibool
page_rec_validate(
/*==============*/
const rec_t* rec, /*!< in: physical record */
const ulint* offsets);/*!< in: array returned by rec_get_offsets() */
#ifdef UNIV_DEBUG
/***************************************************************//**
Checks that the first directory slot points to the infimum record and
the last to the supremum. This function is intended to track if the
bug fixed in 4.0.14 has caused corruption to users' databases. */
void
page_check_dir(
/*===========*/
const page_t* page); /*!< in: index page */
#endif /* UNIV_DEBUG */
/***************************************************************//**
This function checks the consistency of an index page when we do not
know the index. This is also resilient so that this should never crash
even if the page is total garbage.
@return TRUE if ok */
ibool
page_simple_validate_old(
/*=====================*/
const page_t* page); /*!< in: index page in ROW_FORMAT=REDUNDANT */
/***************************************************************//**
This function checks the consistency of an index page when we do not
know the index. This is also resilient so that this should never crash
even if the page is total garbage.
@return TRUE if ok */
ibool
page_simple_validate_new(
/*=====================*/
const page_t* page); /*!< in: index page in ROW_FORMAT!=REDUNDANT */
/***************************************************************//**
This function checks the consistency of an index page.
@return TRUE if ok */
ibool
page_validate(
/*==========*/
const page_t* page, /*!< in: index page */
dict_index_t* index); /*!< in: data dictionary index containing
the page record type definition */
/***************************************************************//**
Looks in the page record list for a record with the given heap number.
@return record, NULL if not found */
const rec_t*
page_find_rec_with_heap_no(
/*=======================*/
const page_t* page, /*!< in: index page */
ulint heap_no);/*!< in: heap number */
2014-05-05 18:20:28 +02:00
/** Get the last non-delete-marked record on a page.
@param[in] page index tree leaf page
@return the last record, not delete-marked
@retval infimum record if all records are delete-marked */
const rec_t*
page_find_rec_max_not_deleted(
const page_t* page);
2015-06-16 10:57:05 +02:00
/** Issue a warning when the checksum that is stored in the page is valid,
but different than the global setting innodb_checksum_algorithm.
@param[in] current_algo current checksum algorithm
@param[in] page_checksum page valid checksum
@param[in] page_id page identifier */
2015-06-16 10:57:05 +02:00
void
page_warn_strict_checksum(
srv_checksum_algorithm_t curr_algo,
srv_checksum_algorithm_t page_checksum,
const page_id_t& page_id);
2015-06-16 10:57:05 +02:00
#ifdef UNIV_MATERIALIZE
#undef UNIV_INLINE
#define UNIV_INLINE UNIV_INLINE_ORIGINAL
#endif
#endif /* !UNIV_INNOCHECKSUM */
#include "page0page.ic"
#endif