Under unknown circumstances, the SQL layer may wrongly disregard an
invocation of thd_mark_transaction_to_rollback() when an InnoDB
transaction had been aborted (rolled back) due to one of the following errors:
* HA_ERR_LOCK_DEADLOCK
* HA_ERR_RECORD_CHANGED (if innodb_snapshot_isolation=ON)
* HA_ERR_LOCK_WAIT_TIMEOUT (if innodb_rollback_on_timeout=ON)
Such an error used to cause a crash of InnoDB during transaction commit.
These changes aim to catch and report the error earlier, so that not only
this crash can be avoided but also the original root cause be found and
fixed more easily later.
The idea of this fix is from Michael 'Monty' Widenius.
HA_ERR_ROLLBACK: A new error code that will be translated into
ER_ROLLBACK_ONLY, signalling that the current transaction
has been aborted and the only allowed action is ROLLBACK.
trx_t::state: Add TRX_STATE_ABORTED that is like
TRX_STATE_NOT_STARTED, but noting that the transaction had been
rolled back and aborted.
trx_t::is_started(): Replaces trx_is_started().
ha_innobase: Check the transaction state in various places.
Simplify the logic around SAVEPOINT.
ha_innobase::is_valid_trx(): Replaces ha_innobase::is_read_only().
The InnoDB logic around transaction savepoints, commit, and rollback
was unnecessarily complex and might have contributed to this
inconsistency. So, we are simplifying that logic as well.
trx_savept_t: Replace with const undo_no_t*. When we rollback to
a savepoint, all we need to know is the number of undo log records
that must survive.
trx_named_savept_t, DB_NO_SAVEPOINT: Remove. We can store undo_no_t
directly in the space allocated at innobase_hton->savepoint_offset.
fts_trx_create(): Do not copy previous savepoints.
fts_savepoint_rollback(): If a savepoint was not found, roll back
everything after the default savepoint of fts_trx_create().
The test innodb_fts.savepoint is extended to cover this code.
Reviewed by: Vladislav Lesin
Tested by: Matthias Leich
- Replace statement fails with duplicate key error when multiple
unique key conflict happens. Reason is that Server expects the
InnoDB engine to store the confliciting keys in ascending order.
But the InnoDB doesn't store the conflicting keys
in ascending order.
Fix:
===
- Enable HA_DUPLICATE_KEY_NOT_IN_ORDER for InnoDB storage engine
only when unique index order is different in .frm and innodb dictionary.
into a separate transaction_participant structure
handlerton inherits it, so handlerton itself doesn't change.
but entities that only need to participate in a transaction,
like binlog or online alter log, use a transaction_participant
and no longer need to pretend to be a full-blown but invisible
storage engine which doesn't support create table.
the information about index algorithm was stored in two
places inconsistently split between both.
BTREE index could have key->algorithm == HA_KEY_ALG_BTREE, if the user
explicitly specified USING BTREE or HA_KEY_ALG_UNDEF, if not.
RTREE index had key->algorithm == HA_KEY_ALG_RTREE
and always had key->flags & HA_SPATIAL
FULLTEXT index had key->algorithm == HA_KEY_ALG_FULLTEXT
and always had key->flags & HA_FULLTEXT
HASH index had key->algorithm == HA_KEY_ALG_HASH or HA_KEY_ALG_UNDEF
long unique index always had key->algorithm == HA_KEY_ALG_LONG_HASH
In this commit:
All indexes except BTREE and HASH always have key->algorithm
set, HA_SPATIAL and HA_FULLTEXT flags are not used anymore (except
for storage to keep frms backward compatible).
As a side effect ALTER TABLE now detects FULLTEXT index renames correctly
The method was declared to return an unsigned integer, but it is
really a boolean (and used as such by all callers).
A secondary change is the addition of "const" and "noexcept" to this
method.
In ha_mroonga.cpp, I also added "inline" to the two helper methods of
referenced_by_foreign_key(). This allows the compiler to flatten the
method.
crash recovery
Summary
=======
When doing server recovery, the active transactions will be rolled
back by InnoDB background rollback thread automatically. The
prepared transactions will be committed or rolled back accordingly
by binlog recovery. Binlog recovery is done in main thread before
the server can provide service to users. If there is a big
transaction to rollback, the server will not available for a long
time.
This patch provides a way to rollback the prepared transactions
asynchronously. Thus the rollback will not block server startup.
Design
======
- Handler::recover_rollback_by_xid()
This patch provides a new handler interface to rollback transactions
in recover phase. InnoDB just set the transaction's state to active.
Then the transaction will be rolled back by the background rollback
thread.
- Handler::signal_tc_log_recover_done()
This function is called after tc log is opened(typically binlog opened)
has done. When this function is called, all transactions will be rolled
back have been reverted to ACTIVE state. Thus it starts rollback thread
to rollback the transactions.
- Background rollback thread
With this patch, background rollback thread is defered to run until binlog
recovery is finished. It is started by innobase_tc_log_recovery_done().
During a Sysbench oltp_point_select workload with 1 table and 400
concurrent connections, a bottleneck on dict_table_t::lock_mutex was
observed in ha_innobase::info_low().
dict_table_t::lock_latch: Replaces lock_mutex.
In ha_innobase::info_low() and several other places, we will acquire
a shared dict_table_t::lock_latch or we may elide the latch if
hardware memory transactions are available.
innobase_build_v_templ(): Remove the parameter "bool locked", and
require the caller to hold exclusive dict_table_t::lock_latch
(instead of holding an exclusive dict_sys.latch).
Tested by: Vladislav Vaintroub
Reviewed by: Vladislav Vaintroub
This task is to ensure we have a clear definition and rules of how to
repair or optimize a table.
The rules are:
- REPAIR should be used with tables that are crashed and are
unreadable (hardware issues with not readable blocks, blocks with
'unexpected data' etc)
- OPTIMIZE table should be used to optimize the storage layout for the
table (recover space for delete rows and optimize the index
structure.
- ALTER TABLE table_name FORCE should be used to rebuild the .frm file
(the table definition) and the table (with the original table row
format). If the table is from and older MariaDB/MySQL release with a
different storage format, it will convert the data to the new
format. ALTER TABLE ... FORCE is used as part of mariadb-upgrade
Here follows some more background:
The 3 ways to repair a table are:
1) ALTER TABLE table_name FORCE" (not other options).
As an alias we allow: "ALTER TABLE table_name ENGINE=original_engine"
2) "REPAIR TABLE" (without FORCE)
3) "OPTIMIZE TABLE"
All of the above commands will optimize row space usage (which means that
space will be needed to hold a temporary copy of the table) and
re-generate all indexes. They will also try to replicate the original
table definition as exact as possible.
For ALTER TABLE and "REPAIR TABLE without FORCE", the following will hold:
If the table is from an older MariaDB version and data conversion is
needed (for example for old type HASH columns, MySQL JSON type or new
TIMESTAMP format) "ALTER TABLE table_name FORCE, algorithm=COPY" will be
used.
The differences between the algorithms are
1) Will use the fastest algorithm the engine supports to do a full repair
of the table (except if data conversions are is needed).
2) Will use the storage engine internal REPAIR facility (MyISAM, Aria).
If the engine does not support REPAIR then
"ALTER TABLE FORCE, ALGORITHM=COPY" will be used.
If there was data incompatibilities (which means that FORCE was used)
then there will be a warning after REPAIR that ALTER TABLE FORCE is
still needed.
The reason for this is that REPAIR may be able to go around data
errors (wrong incompatible data, crashed or unreadable sectors) that
ALTER TABLE cannot do.
3) Will use the storage engine internal OPTIMIZE. If engine does not
support optimize, then "ALTER TABLE FORCE" is used.
The above will ensure that ALTER TABLE FORCE is able to
correct almost any errors in the row or index data. In case of
corrupted blocks then REPAIR possible followed by ALTER TABLE is needed.
This is important as mariadb-upgrade executes ALTER TABLE table_name
FORCE for any table that must be re-created.
Bugs fixed with InnoDB tables when using ALTER TABLE FORCE:
- No error for INNODB_DEFAULT_ROW_FORMAT=COMPACT even if row length
would be too wide. (Independent of innodb_strict_mode).
- Tables using symlinks will be symlinked after any of the above commands
(Independent of the setting of --symbolic-links)
If one specifies an algorithm together with ALTER TABLE FORCE, things
will work as before (except if data conversion is required as then
the COPY algorithm is enforced).
ALTER TABLE .. OPTIMIZE ALL PARTITIONS will work as before.
Other things:
- FORCE argument added to REPAIR to allow one to first run internal
repair to fix damaged blocks and then follow it with ALTER TABLE.
- REPAIR will not update frm_version if ha_check_for_upgrade() finds
that table is still incompatible with current version. In this case the
REPAIR will end with an error.
- REPAIR for storage engines that does not have native repair, like InnoDB,
is now using ALTER TABLE FORCE.
- REPAIR csv-table USE_FRM now works.
- It did not work before as CSV tables had extension list in wrong
order.
- Default error messages length for %M increased from 128 to 256 to not
cut information from REPAIR.
- Documented HA_ADMIN_XX variables related to repair.
- Added HA_ADMIN_NEEDS_DATA_CONVERSION to signal that we have to
do data conversions when converting the table (and thus ALTER TABLE
copy algorithm is needed).
- Fixed typo in error message (caused test changes).
(Based on original patch by Oleksandr Byelkin)
Multi-table DELETE can execute via "buffered" mode: at phase #1 it collects
rowids of rows to be deleted, then at phase #2 in multi_delete::do_deletes()
it calls handler->rnd_pos() to read rows to be deleted and deletes them.
The problem occurred when phase #1 used Rowid Filter on the table that
phase #2 would be deleting from.
In InnoDB, h->rnd_init(scan=false) and h->rnd_pos() is an index scan over PK
under the hood. So, at phase #2 ha_innobase::rnd_init() would try to use the
Rowid Filter and hit an assertion inside ha_innobase::rnd_init().
Note that multi-table UPDATE works similarly but was not affected, because
patch for MDEV-7487 added code to disable rowid filter for phase #2 in
multi_update::do_updates().
This patch changes the approach:
- It makes InnoDB not use Rowid Filter in rnd_pos() scans: it is disabled in
ha_innobase::rnd_init() and enabled back in ha_innobase::rnd_end().
- multi_update::do_updates() no longer disables Rowid Filter for phase#2 as
it is no longer necessary.
This patch also fixes:
MDEV-33050 Build-in schemas like oracle_schema are accent insensitive
MDEV-33084 LASTVAL(t1) and LASTVAL(T1) do not work well with lower-case-table-names=0
MDEV-33085 Tables T1 and t1 do not work well with ENGINE=CSV and lower-case-table-names=0
MDEV-33086 SHOW OPEN TABLES IN DB1 -- is case insensitive with lower-case-table-names=0
MDEV-33088 Cannot create triggers in the database `MYSQL`
MDEV-33103 LOCK TABLE t1 AS t2 -- alias is not case sensitive with lower-case-table-names=0
MDEV-33109 DROP DATABASE MYSQL -- does not drop SP with lower-case-table-names=0
MDEV-33110 HANDLER commands are case insensitive with lower-case-table-names=0
MDEV-33119 User is case insensitive in INFORMATION_SCHEMA.VIEWS
MDEV-33120 System log table names are case insensitive with lower-cast-table-names=0
- Removing the virtual function strnncoll() from MY_COLLATION_HANDLER
- Adding a wrapper function CHARSET_INFO::streq(), to compare
two strings for equality. For now it calls strnncoll() internally.
In the future it will turn into a virtual function.
- Adding new accent sensitive case insensitive collations:
- utf8mb4_general1400_as_ci
- utf8mb3_general1400_as_ci
They implement accent sensitive case insensitive comparison.
The weight of a character is equal to the code point of its
upper case variant. These collations use Unicode-14.0.0 casefolding data.
The result of
my_charset_utf8mb3_general1400_as_ci.strcoll()
is very close to the former
my_charset_utf8mb3_general_ci.strcasecmp()
There is only a difference in a couple dozen rare characters, because:
- the switch from "tolower" to "toupper" comparison, to make
utf8mb3_general1400_as_ci closer to utf8mb3_general_ci
- the switch from Unicode-3.0.0 to Unicode-14.0.0
This difference should be tolarable. See the list of affected
characters in the MDEV description.
Note, utf8mb4_general1400_as_ci correctly handles non-BMP characters!
Unlike utf8mb4_general_ci, it does not treat all BMP characters
as equal.
- Adding classes representing names of the file based database objects:
Lex_ident_db
Lex_ident_table
Lex_ident_trigger
Their comparison collation depends on the underlying
file system case sensitivity and on --lower-case-table-names
and can be either my_charset_bin or my_charset_utf8mb3_general1400_as_ci.
- Adding classes representing names of other database objects,
whose names have case insensitive comparison style,
using my_charset_utf8mb3_general1400_as_ci:
Lex_ident_column
Lex_ident_sys_var
Lex_ident_user_var
Lex_ident_sp_var
Lex_ident_ps
Lex_ident_i_s_table
Lex_ident_window
Lex_ident_func
Lex_ident_partition
Lex_ident_with_element
Lex_ident_rpl_filter
Lex_ident_master_info
Lex_ident_host
Lex_ident_locale
Lex_ident_plugin
Lex_ident_engine
Lex_ident_server
Lex_ident_savepoint
Lex_ident_charset
engine_option_value::Name
- All the mentioned Lex_ident_xxx classes implement a method streq():
if (ident1.streq(ident2))
do_equal();
This method works as a wrapper for CHARSET_INFO::streq().
- Changing a lot of "LEX_CSTRING name" to "Lex_ident_xxx name"
in class members and in function/method parameters.
- Replacing all calls like
system_charset_info->coll->strcasecmp(ident1, ident2)
to
ident1.streq(ident2)
- Taking advantage of the c++11 user defined literal operator
for LEX_CSTRING (see m_strings.h) and Lex_ident_xxx (see lex_ident.h)
data types. Use example:
const Lex_ident_column primary_key_name= "PRIMARY"_Lex_ident_column;
is now a shorter version of:
const Lex_ident_column primary_key_name=
Lex_ident_column({STRING_WITH_LEN("PRIMARY")});
Under terms of MDEV 27490 we'll add support for non-BMP identifiers
and upgrade casefolding information to Unicode version 14.0.0.
In Unicode-14.0.0 conversion to lower and upper cases can increase octet length
of the string, so conversion won't be possible in-place any more.
This patch removes virtual functions performing in-place casefolding:
- my_charset_handler_st::casedn_str()
- my_charset_handler_st::caseup_str()
and fixes the code to use the non-inplace functions instead:
- my_charset_handler_st::casedn()
- my_charset_handler_st::caseup()
MDEV-33308 CHECK TABLE is modifying .frm file even if --read-only
As noted in commit d0ef1aaf61,
MySQL as well as older versions of MariaDB server would during
ALTER TABLE ... IMPORT TABLESPACE write bogus values to the
PAGE_MAX_TRX_ID field to pages of the clustered index, instead of
letting that field remain 0.
In commit 8777458a6e this field
was repurposed for PAGE_ROOT_AUTO_INC in the clustered index root page.
To avoid trouble when upgrading from MySQL or older versions of MariaDB,
we will try to detect and correct bogus values of PAGE_ROOT_AUTO_INC
when opening a table for the first time from the SQL layer.
btr_read_autoinc_with_fallback(): Add the parameters to mysql_version,max
to indicate the TABLE_SHARE::mysql_version of the .frm file and the
maximum value allowed for the type of the AUTO_INCREMENT column.
In case the table was originally created in MySQL or an older version of
MariaDB, read also the maximum value of the AUTO_INCREMENT column from
the table and reset the PAGE_ROOT_AUTO_INC if it is above the limit.
dict_table_t::get_index(const dict_col_t &) const: Find an index that
starts with the specified column.
ha_innobase::check_for_upgrade(): Return HA_ADMIN_FAILED if InnoDB
needs upgrading but is in read-only mode. In this way, the call to
update_frm_version() will be skipped.
row_import_autoinc(): Adjust the AUTO_INCREMENT column at the end of
ALTER TABLE...IMPORT TABLESPACE. This refinement was suggested by
Debarun Banerjee.
The changes outside InnoDB were developed by Michael 'Monty' Widenius:
Added print_check_msg() service for easy reporting of check/repair messages
in ENGINE=Aria and ENGINE=InnoDB.
Fixed that CHECK TABLE do not update the .frm file under --read-only.
Added 'handler_flags' to HA_CHECK_OPT as a way for storage engines to
store state from handler::check_for_upgrade().
Reviewed by: Debarun Banerjee
ht->start_consistent_snapshot() is also not a way,
because some engines (e.g. rocksdb) only do it readonly.
instead, downgrade the lock after reading the first row
(which implicitly opens a read view).
* Log rows in online_alter_binlog.
* Table online data is replicated within dedicated binlog file
* Cached data is written on commit.
* Versioning is fully supported.
* Works both wit and without binlog enabled.
* For now savepoints setup is forbidden while ONLINE ALTER goes on.
Extra support is required. We can simply log the SAVEPOINT query events
and replicate them together with row events. But it's not implemented
for now.
* Cache flipping:
We want to care for the possible bottleneck in the online alter binlog
reading/writing in advance.
IO_CACHE does not provide anything better that sequential access,
besides, only a single write is mutex-protected, which is not suitable,
since we should write a transaction atomically.
To solve this, a special layer on top Event_log is implemented.
There are two IO_CACHE files underneath: one for reading, and one for
writing.
Once the read cache is empty, an exclusive lock is acquired (we can wait
for a currently active transaction finish writing), and flip() is emitted,
i.e. the write cache is reopened for read, and the read cache is emptied,
and reopened for writing.
This reminds a buffer flip that happens in accelerated graphics
(DirectX/OpenGL/etc).
Cache_flip_event_log is considered non-blocking for a single reader and a
single writer in this sense, with the only lock held by reader during flip.
An alternative approach by implementing a fair concurrent circular buffer
is described in MDEV-24676.
* Cache managers:
We have two cache sinks: statement and transactional.
It is important that the changes are first cached per-statement and
per-transaction.
If a statement fails, then only statement data is rolled back. The
transaction moves along, however.
Turns out, there's no guarantee that TABLE well persist in
thd->open_tables to the transaction commit moment.
If an error occurs, tables from statement are purged.
Therefore, we can't store te caches in TABLE. Ideally, it should be
handlerton, but we cut the corner and store it in THD in a list.
Allow ALTER TABLE ... IMPORT TABLESPACE without creating the table
followed by discarding the tablespace.
That is, assuming we want to import table t1 to t2, instead of
CREATE TABLE t2 LIKE t1;
ALTER TABLE t2 DISCARD TABLESPACE;
FLUSH TABLES t1 FOR EXPORT;
--copy_file $MYSQLD_DATADIR/test/t1.cfg $MYSQLD_DATADIR/test/t2.cfg
--copy_file $MYSQLD_DATADIR/test/t1.ibd $MYSQLD_DATADIR/test/t2.ibd
UNLOCK TABLES;
ALTER TABLE t2 IMPORT TABLESPACE;
We can simply do
FLUSH TABLES t1 FOR EXPORT;
--copy_file $MYSQLD_DATADIR/test/t1.cfg $MYSQLD_DATADIR/test/t2.cfg
--copy_file $MYSQLD_DATADIR/test/t1.frm $MYSQLD_DATADIR/test/t2.frm
--copy_file $MYSQLD_DATADIR/test/t1.ibd $MYSQLD_DATADIR/test/t2.ibd
UNLOCK TABLES;
ALTER TABLE t2 IMPORT TABLESPACE;
We achieve this by creating a "stub" table in the second scenario
while opening the table, where t2 does not exist but needs to import
from t1. The "stub" table is similar to a table that is created but
then instructed to discard its tablespace.
We include tests with various row formats, encryption, with indexes
and auto-increment.
The deprecated parameters will be removed:
innodb_defragment
innodb_defragment_n_pages
innodb_defragment_stats_accuracy
innodb_defragment_fill_factor_n_recs
innodb_defragment_fill_factor
innodb_defragment_frequency
The mysql.innodb_index_stats.stat_name values 'n_page_split' and
'n_pages_freed' will lose their special meaning.
The related changes to OPTIMIZE TABLE in InnoDB will be removed as well.
The parameter innodb_optimize_fulltext_only will retain its special
meaning in OPTIMIZE TABLE.
Tested by: Matthias Leich
The issue was that when limit is used,
SQL_SELECT::test_quick_select would set the cost of table scan to be
unreasonable high to force a range to be used.
The problem with this approach was that range was used even when the
cost of range, when it would only read 'limit rows' would be higher
than the cost of a table scan.
This patch fixes it by not accepting ranges when the range can never
have a lower cost than a table scan, even if every row would match the
WHERE clause.
This makes it easier to compare different costs and also allows
the optimizer to optimizer different storage engines more reliably.
- Added tests/check_costs.pl, a tool to verify optimizer cost calculations.
- Most engine costs has been found with this program. All steps to
calculate the new costs are documented in Docs/optimizer_costs.txt
- User optimizer_cost variables are given in microseconds (as individual
costs can be very small). Internally they are stored in ms.
- Changed DISK_READ_COST (was DISK_SEEK_BASE_COST) from a hard disk cost
(9 ms) to common SSD cost (400MB/sec).
- Removed cost calculations for hard disks (rotation etc).
- Changed the following handler functions to return IO_AND_CPU_COST.
This makes it easy to apply different cost modifiers in ha_..time()
functions for io and cpu costs.
- scan_time()
- rnd_pos_time() & rnd_pos_call_time()
- keyread_time()
- Enhanched keyread_time() to calculate the full cost of reading of a set
of keys with a given number of ranges and optional number of blocks that
need to be accessed.
- Removed read_time() as keyread_time() + rnd_pos_time() can do the same
thing and more.
- Tuned cost for: heap, myisam, Aria, InnoDB, archive and MyRocks.
Used heap table costs for json_table. The rest are using default engine
costs.
- Added the following new optimizer variables:
- optimizer_disk_read_ratio
- optimizer_disk_read_cost
- optimizer_key_lookup_cost
- optimizer_row_lookup_cost
- optimizer_row_next_find_cost
- optimizer_scan_cost
- Moved all engine specific cost to OPTIMIZER_COSTS structure.
- Changed costs to use 'records_out' instead of 'records_read' when
recalculating costs.
- Split optimizer_costs.h to optimizer_costs.h and optimizer_defaults.h.
This allows one to change costs without having to compile a lot of
files.
- Updated costs for filter lookup.
- Use a better cost estimate in best_extension_by_limited_search()
for the sorting cost.
- Fixed previous issues with 'filtered' explain column as we are now
using 'records_out' (min rows seen for table) to calculate filtering.
This greatly simplifies the filtering code in
JOIN_TAB::save_explain_data().
This change caused a lot of queries to be optimized differently than
before, which exposed different issues in the optimizer that needs to
be fixed. These fixes are in the following commits. To not have to
change the same test case over and over again, the changes in the test
cases are done in a single commit after all the critical change sets
are done.
InnoDB changes:
- Updated InnoDB to not divide big range cost with 2.
- Added cost for InnoDB (innobase_update_optimizer_costs()).
- Don't mark clustered primary key with HA_KEYREAD_ONLY. This will
prevent that the optimizer is trying to use index-only scans on
the clustered key.
- Disabled ha_innobase::scan_time() and ha_innobase::read_time() and
ha_innobase::rnd_pos_time() as the default engine cost functions now
works good for InnoDB.
Other things:
- Added --show-query-costs (\Q) option to mysql.cc to show the query
cost after each query (good when working with query costs).
- Extended my_getopt with GET_ADJUSTED_VALUE which allows one to adjust
the value that user is given. This is used to change cost from
microseconds (user input) to milliseconds (what the server is
internally using).
- Added include/my_tracker.h ; Useful include file to quickly test
costs of a function.
- Use handler::set_table() in all places instead of 'table= arg'.
- Added SHOW_OPTIMIZER_COSTS to sys variables. These are input and
shown in microseconds for the user but stored as milliseconds.
This is to make the numbers easier to read for the user (less
pre-zeros). Implemented in 'Sys_var_optimizer_cost' class.
- In test_quick_select() do not use index scans if 'no_keyread' is set
for the table. This is what we do in other places of the server.
- Added THD parameter to Unique::get_use_cost() and
check_index_intersect_extension() and similar functions to be able
to provide costs to called functions.
- Changed 'records' to 'rows' in optimizer_trace.
- Write more information to optimizer_trace.
- Added INDEX_BLOCK_FILL_FACTOR_MUL (4) and INDEX_BLOCK_FILL_FACTOR_DIV (3)
to calculate usage space of keys in b-trees. (Before we used numeric
constants).
- Removed code that assumed that b-trees has similar costs as binary
trees. Replaced with engine calls that returns the cost.
- Added Bitmap::find_first_bit()
- Added timings to join_cache for ANALYZE table (patch by Sergei Petrunia).
- Added records_init and records_after_filter to POSITION to remember
more of what best_access_patch() calculates.
- table_after_join_selectivity() changed to recalculate 'records_out'
based on the new fields from best_access_patch()
Bug fixes:
- Some queries did not update last_query_cost (was 0). Fixed by moving
setting thd->...last_query_cost in JOIN::optimize().
- Write '0' as number of rows for const tables with a matching row.
Some internals:
- Engine cost are stored in OPTIMIZER_COSTS structure. When a
handlerton is created, we also created a new cost variable for the
handlerton. We also create a new variable if the user changes a
optimizer cost for a not yet loaded handlerton either with command
line arguments or with SET
@@global.engine.optimizer_cost_variable=xx.
- There are 3 global OPTIMIZER_COSTS variables:
default_optimizer_costs The default costs + changes from the
command line without an engine specifier.
heap_optimizer_costs Heap table costs, used for temporary tables
tmp_table_optimizer_costs The cost for the default on disk internal
temporary table (MyISAM or Aria)
- The engine cost for a table is stored in table_share. To speed up
accesses the handler has a pointer to this. The cost is copied
to the table on first access. If one wants to change the cost one
must first update the global engine cost and then do a FLUSH TABLES.
This was done to be able to access the costs for an open table
without any locks.
- When a handlerton is created, the cost are updated the following way:
See sql/keycaches.cc for details:
- Use 'default_optimizer_costs' as a base
- Call hton->update_optimizer_costs() to override with the engines
default costs.
- Override the costs that the user has specified for the engine.
- One handler open, copy the engine cost from handlerton to TABLE_SHARE.
- Call handler::update_optimizer_costs() to allow the engine to update
cost for this particular table.
- There are two costs stored in THD. These are copied to the handler
when the table is used in a query:
- optimizer_where_cost
- optimizer_scan_setup_cost
- Simply code in best_access_path() by storing all cost result in a
structure. (Idea/Suggestion by Igor)
Before this patch, when calculating the cost of fetching and using a
row/key from the engine, we took into account the cost of finding a
row or key from the engine, but did not consistently take into account
index only accessed, clustered key or covered keys for all access
paths.
The cost of the WHERE clause (TIME_FOR_COMPARE) was not consistently
considered in best_access_path(). TIME_FOR_COMPARE was used in
calculation in other places, like greedy_search(), but was in some
cases (like scans) done an a different number of rows than was
accessed.
The cost calculation of row and index scans didn't take into account
the number of rows that where accessed, only the number of accepted
rows.
When using a filter, the cost of index_only_reads and cost of
accessing and disregarding 'filtered rows' where not taken into
account, which made filters cost less than there actually where.
To remedy the above, the following key & row fetch related costs
has been added:
- The cost of fetching and using a row is now split into different costs:
- key + Row fetch cost (as before) but multiplied with the variable
'optimizer_cache_cost' (default to 0.5). This allows the user to
tell the optimizer the likehood of finding the key and row in the
engine cache.
- ROW_COPY_COST, The cost copying a row from the engine to the
sql layer or creating a row from the join_cache to the record
buffer. Mostly affects table scan costs.
- ROW_LOOKUP_COST, the cost of fetching a row by rowid.
- KEY_COPY_COST the cost of finding the next key and copying it from
the engine to the SQL layer. This is used when we calculate the cost
index only reads. It makes index scans more expensive than before if
they cover a lot of rows. (main.index_merge_myisam)
- KEY_LOOKUP_COST, the cost of finding the first key in a range.
This replaces the old define IDX_LOOKUP_COST, but with a higher cost.
- KEY_NEXT_FIND_COST, the cost of finding the next key (and rowid).
when doing a index scan and comparing the rowid to the filter.
Before this cost was assumed to be 0.
All of the above constants/variables are now tuned to be somewhat in
proportion of executing complexity to each other. There is tuning
need for these in the future, but that can wait until the above are
made user variables as that will make tuning much easier.
To make the usage of the above easy, there are new (not virtual)
cost calclation functions in handler:
- ha_read_time(), like read_time(), but take optimizer_cache_cost into
account.
- ha_read_and_copy_time(), like ha_read_time() but take into account
ROW_COPY_TIME
- ha_read_and_compare_time(), like ha_read_and_copy_time() but take
TIME_FOR_COMPARE into account.
- ha_rnd_pos_time(). Read row with row id, taking ROW_COPY_COST
into account. This is used with filesort where we don't need
to execute the WHERE clause again.
- ha_keyread_time(), like keyread_time() but take
optimizer_cache_cost into account.
- ha_keyread_and_copy_time(), like ha_keyread_time(), but add
KEY_COPY_COST.
- ha_key_scan_time(), like key_scan_time() but take
optimizer_cache_cost nto account.
- ha_key_scan_and_compare_time(), like ha_key_scan_time(), but add
KEY_COPY_COST & TIME_FOR_COMPARE.
I also added some setup costs for doing different types of scans and
creating temporary tables (on disk and in memory). This encourages
the optimizer to not use these for simple 'a few row' lookups if
there are adequate key lookup strategies.
- TABLE_SCAN_SETUP_COST, cost of starting a table scan.
- INDEX_SCAN_SETUP_COST, cost of starting an index scan.
- HEAP_TEMPTABLE_CREATE_COST, cost of creating in memory
temporary table.
- DISK_TEMPTABLE_CREATE_COST, cost of creating an on disk temporary
table.
When calculating cost of fetching ranges, we had a cost of
IDX_LOOKUP_COST (0.125) for doing a key div for a new range. This is
now replaced with 'io_cost * KEY_LOOKUP_COST (1.0) *
optimizer_cache_cost', which matches the cost we use for 'ref' and
other key lookups. The effect is that the cost is now a bit higher
when we have many ranges for a key.
Allmost all calculation with TIME_FOR_COMPARE is now done in
best_access_path(). 'JOIN::read_time' now includes the full
cost for finding the rows in the table.
In the result files, many of the changes are now again close to what
they where before the "Update cost for hash and cached joins" commit,
as that commit didn't fix the filter cost (too complex to do
everything in one commit).
The above changes showed a lot of a lot of inconsistencies in
optimizer cost calculation. The main objective with the other changes
was to do calculation as similar (and accurate) as possible and to make
different plans more comparable.
Detailed list of changes:
- Calculate index_only_cost consistently and correctly for all scan
and ref accesses. The row fetch_cost and index_only_cost now
takes into account clustered keys, covered keys and index
only accesses.
- cost_for_index_read now returns both full cost and index_only_cost
- Fixed cost calculation of get_sweep_read_cost() to match other
similar costs. This is bases on the assumption that data is more
often stored on SSD than a hard disk.
- Replaced constant 2.0 with new define TABLE_SCAN_SETUP_COST.
- Some scan cost estimates did not take into account
TIME_FOR_COMPARE. Now all scan costs takes this into
account. (main.show_explain)
- Added session variable optimizer_cache_hit_ratio (default 50%). By
adjusting this on can reduce or increase the cost of index or direct
record lookups. The effect of the default is that key lookups is now
a bit cheaper than before. See usage of 'optimizer_cache_cost' in
handler.h.
- JOIN_TAB::scan_time() did not take into account index only scans,
which produced a wrong cost when index scan was used. Changed
JOIN_TAB:::scan_time() to take into consideration clustered and
covered keys. The values are now cached and we only have to call
this function once. Other calls are changed to use the cached
values. Function renamed to JOIN_TAB::estimate_scan_time().
- Fixed that most index cost calculations are done the same way and
more close to 'range' calculations. The cost is now lower than
before for small data sets and higher for large data sets as we take
into account how many keys are read (main.opt_trace_selectivity,
main.limit_rows_examined).
- Ensured that index_scan_cost() ==
range(scan_of_all_rows_in_table_using_one_range) +
MULTI_RANGE_READ_INFO_CONST. One effect of this is that if there
is choice of doing a full index scan and a range-index scan over
almost the whole table then index scan will be preferred (no
range-read setup cost). (innodb.innodb, main.show_explain,
main.range)
- Fixed the EQ_REF and REF takes into account clustered and covered
keys. This changes some plans to use covered or clustered indexes
as these are much cheaper. (main.subselect_mat_cost,
main.state_tables_innodb, main.limit_rows_examined)
- Rowid filter setup cost and filter compare cost now takes into
account fetching and checking the rowid (KEY_NEXT_FIND_COST).
(main.partition_pruning heap.heap_btree main.log_state)
- Added KEY_NEXT_FIND_COST to
Range_rowid_filter_cost_info::lookup_cost to account of the time
to find and check the next key value against the container
- Introduced ha_keyread_time(rows) that takes into account finding
the next row and copying the key value to 'record'
(KEY_COPY_COST).
- Introduced ha_key_scan_time() for calculating an index scan over
all rows.
- Added IDX_LOOKUP_COST to keyread_time() as a startup cost.
- Added index_only_fetch_cost() as a convenience function to
OPT_RANGE.
- keyread_time() cost is slightly reduced to prefer shorter keys.
(main.index_merge_myisam)
- All of the above caused some index_merge combinations to be
rejected because of cost (main.index_intersect). In some cases
'ref' where replaced with index_merge because of the low
cost calculation of get_sweep_read_cost().
- Some index usage moved from PRIMARY to a covering index.
(main.subselect_innodb)
- Changed cost calculation of filter to take KEY_LOOKUP_COST and
TIME_FOR_COMPARE into account. See sql_select.cc::apply_filter().
filter parameters and costs are now written to optimizer_trace.
- Don't use matchings_records_in_range() to try to estimate the number
of filtered rows for ranges. The reason is that we want to ensure
that 'range' is calculated similar to 'ref'. There is also more work
needed to calculate the selectivity when using ranges and ranges and
filtering. This causes filtering column in EXPLAIN EXTENDED to be
100.00 for some cases where range cannot use filtering.
(main.rowid_filter)
- Introduced ha_scan_time() that takes into account the CPU cost of
finding the next row and copying the row from the engine to
'record'. This causes costs of table scan to slightly increase and
some test to changed their plan from ALL to RANGE or ALL to ref.
(innodb.innodb_mysql, main.select_pkeycache)
In a few cases where scan time of very small tables have lower cost
than a ref or range, things changed from ref/range to ALL.
(main.myisam, main.func_group, main.limit_rows_examined,
main.subselect2)
- Introduced ha_scan_and_compare_time() which is like ha_scan_time()
but also adds the cost of the where clause (TIME_FOR_COMPARE).
- Added small cost for creating temporary table for
materialization. This causes some very small tables to use scan
instead of materialization.
- Added checking of the WHERE clause (TIME_FOR_COMPARE) of the
accepted rows to ROR costs in get_best_ror_intersect()
- Removed '- 0.001' from 'join->best_read' and optimize_straight_join()
to ensure that the 'Last_query_cost' status variable contains the
same value as the one that was calculated by the optimizer.
- Take avg_io_cost() into account in handler::keyread_time() and
handler::read_time(). This should have no effect as it's 1.0 by
default, except for heap that overrides these functions.
- Some 'ref_or_null' accesses changed to 'range' because of cost
adjustments (main.order_by)
- Added scan type "scan_with_join_cache" for optimizer_trace. This is
just to show in the trace what kind of scan was used.
- When using 'scan_with_join_cache' take into account number of
preceding tables (as have to restore all fields for all previous
table combination when checking the where clause)
The new cost added is:
(row_combinations * ROW_COPY_COST * number_of_cached_tables).
This increases the cost of join buffering in proportion of the
number of tables in the join buffer. One effect is that full scans
are now done earlier as the cost is then smaller.
(main.join_outer_innodb, main.greedy_optimizer)
- Removed the usage of 'worst_seeks' in cost_for_index_read as it
caused wrong plans to be created; It prefered JT_EQ_REF even if it
would be much more expensive than a full table scan. A related
issue was that worst_seeks only applied to full lookup, not to
clustered or index only lookups, which is not consistent. This
caused some plans to use index scan instead of eq_ref (main.union)
- Changed federated block size from 4096 to 1500, which is the
typical size of an IO packet.
- Added costs for reading rows to Federated. Needed as there is no
caching of rows in the federated engine.
- Added ha_innobase::rnd_pos_time() cost function.
- A lot of extra things added to optimizer trace
- More costs, especially for materialization and index_merge.
- Make lables more uniform
- Fixed a lot of minor bugs
- Added 'trace_started()' around a lot of trace blocks.
- When calculating ORDER BY with LIMIT cost for using an index
the cost did not take into account the number of row retrivals
that has to be done or the cost of comparing the rows with the
WHERE clause. The cost calculated would be just a fraction of
the real cost. Now we calculate the cost as we do for ranges
and 'ref'.
- 'Using index for group-by' is used a bit more than before as
now take into account the WHERE clause cost when comparing
with 'ref' and prefer the method with fewer row combinations.
(main.group_min_max).
Bugs fixed:
- Fixed that we don't calculate TIME_FOR_COMPARE twice for some plans,
like in optimize_straight_join() and greedy_search()
- Fixed bug in save_explain_data where we could test for the wrong
index when displaying 'Using index'. This caused some old plans to
show 'Using index'. (main.subselect_innodb, main.subselect2)
- Fixed bug in get_best_ror_intersect() where 'min_cost' was not
updated, and the cost we compared with was not the one that was
used.
- Fixed very wrong cost calculation for priority queues in
check_if_pq_applicable(). (main.order_by now correctly uses priority
queue)
- When calculating cost of EQ_REF or REF, we added the cost of
comparing the WHERE clause with the found rows, not all row
combinations. This made ref and eq_ref to be regarded way to cheap
compared to other access methods.
- FORCE INDEX cost calculation didn't take into account clustered or
covered indexes.
- JT_EQ_REF cost was estimated as avg_io_cost(), which is half the
cost of a JT_REF key. This may be true for InnoDB primary key, but
not for other unique keys or other engines. Now we use handler
function to calculate the cost, which allows us to handle
consistently clustered, covered keys and not covered keys.
- ha_start_keyread() didn't call extra_opt() if keyread was already
enabled but still changed the 'keyread' variable (which is wrong).
Fixed by not doing anything if keyread is already enabled.
- multi_range_read_info_cost() didn't take into account io_cost when
calculating the cost of ranges.
- fix_semijoin_strategies_for_picked_join_order() used the wrong
record_count when calling best_access_path() for SJ_OPT_FIRST_MATCH
and SJ_OPT_LOOSE_SCAN.
- Hash joins didn't provide correct best_cost to the upper level, which
means that the cost for hash_joins more expensive than calculated
in best_access_path (a difference of 10x * TIME_OF_COMPARE).
This is fixed in the new code thanks to that we now include
TIME_OF_COMPARE cost in 'read_time'.
Other things:
- Added some 'if (thd->trace_started())' to speed up code
- Removed not used function Cost_estimate::is_zero()
- Simplified testing of HA_POS_ERROR in get_best_ror_intersect().
(No cost changes)
- Moved ha_start_keyread() from join_read_const_table() to join_read_const()
to enable keyread for all types of JT_CONST tables.
- Made a few very short functions inline in handler.h
Notes:
- In main.rowid_filter the join order of order and lineitem is swapped.
This is because the cost of doing a range fetch of lineitem(98 rows) is
almost as big as the whole join of order,lineitem. The filtering will
also ensure that we only have to do very small key fetches of the rows
in lineitem.
- main.index_merge_myisam had a few changes where we are now using
less keys for index_merge. This is because index scans are now more
expensive than before.
- handler->optimizer_cache_cost is updated in ha_external_lock().
This ensures that it is up to date per statements.
Not an optimal solution (for locked tables), but should be ok for now.
- 'DELETE FROM t1 WHERE t1.a > 0 ORDER BY t1.a' does not take cost of
filesort into consideration when table scan is chosen.
(main.myisam_explain_non_select_all)
- perfschema.table_aggregate_global_* has changed because an update
on a table with 1 row will now use table scan instead of key lookup.
TODO in upcomming commits:
- Fix selectivity calculation for ranges with and without filtering and
when there is a ref access but scan is chosen.
For this we have to store the lowest known value for
'accepted_records' in the OPT_RANGE structure.
- Change that records_read does not include filtered rows.
- test_if_cheaper_ordering() needs to be updated to properly calculate
costs. This will fix tests like main.order_by_innodb,
main.single_delete_update
- Extend get_range_limit_read_cost() to take into considering
cost_for_index_read() if there where no quick keys. This will reduce
the computed cost for ORDER BY with LIMIT in some cases.
(main.innodb_ext_key)
- Fix that we take into account selectivity when counting the number
of rows we have to read when considering using a index table scan to
resolve ORDER BY.
- Add new calculation for rnd_pos_time() where we take into account the
benefit of reading multiple rows from the same page.
ha_innobase::referenced_by_foreign_key(): Protect the check with
dict_sys.freeze(), to prevent races with TRUNCATE TABLE.
The test innodb.instant_alter_crash has been adjusted for this
additional locking.
dict_table_is_referenced_by_foreign_key(): Removed (merged to
the only caller).
create_table_info_t::create_table(): Ignore missing indexes for
FOREIGN KEY constraints if foreign_key_checks=0.
create_table_info_t::create_table_update_dict(): Rewritten as
a static function. Do not return any error.
ha_innobase::create(): When trx!=nullptr and we are operating
on a persistent table, do not rollback, commit, or release the
data dictionary latch.
ha_innobase::truncate(): Protect the entire critical section
with an exclusive dict_sys.latch, so that
ha_innobase::referenced_by_foreign_key() on referenced tables
will return a consistent result. In case of a failure,
invoke dict_load_foreigns() to restore also any FOREIGN KEY
constraints.
ha_innobase::free_foreign_key_create_info(): Define inline.
lock_release(): Disregard innodb_evict_tables_on_commit_debug=ON
when dict_sys.locked() holds. It would hold when fts_load_stopword()
is invoked by create_table_info_t::create_table_update_dict().
dict_sys_t::locked(): Return whether the current thread is holding
the exclusive dict_sys.latch.
dict_sys_t::frozen_not_locked(): Return whether any thread is
holding a shared dict_sys.latch.
In the test main.mysql_upgrade, the InnoDB persistent statistics
will no longer be recalculated in ha_innobase::open() as part of
CHECK TABLE ... FOR UPGRADE. They were deleted earlier in the test.
Tested by: Matthias Leich
Making changes to wsrep_mysqld.h causes large parts of server code to
be recompiled. The reason is that wsrep_mysqld.h is included by
sql_class.h, even tough very little of wsrep_mysqld.h is needed in
sql_class.h. This commit introduces a new header file, wsrep_on.h,
which is meant to be included from sql_class.h, and contains only
macros and variable declarations used to determine whether wsrep is
enabled.
Also, header wsrep.h should only contain definitions that are also
used outside of sql/. Therefore, move WSREP_TO_ISOLATION* and
WSREP_SYNC_WAIT macros to wsrep_mysqld.h.
Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>
