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
ha_storage_put_memlim(): Invoke my_crc32c() to "fold", and traverse
the hash table only once.
fold_lock(): Remove some redundant conditions and use my_crc32c()
instead of ut_fold_ulint_pair().
trx_i_s_cache_t::add(): Replaces add_lock_to_cache(),
search_innodb_locks(), and locks_row_eq_lock(). Avoid duplicated
traversal of the hash table.
Reviewed by: Debarun Banerjee
Since commit fb335b48b5 we may have
a null pointer in purge_sys.query when fetch_data_into_cache() is
invoked and innodb_force_recovery>4. This is because the call to
purge_sys.create() would be skipped.
fetch_data_into_cache(): Load the purge_sys pseudo transaction pointer
to a local variable (null pointer if purge_sys is not initialized).
trx_t::will_lock: Changed the type to bool.
trx_t::is_autocommit_non_locking(): Replaces
trx_is_autocommit_non_locking().
trx_is_ac_nl_ro(): Remove (replaced with equivalent assertion expressions).
assert_trx_nonlocking_or_in_list(): Remove.
Replaced with at least as strict checks in each place.
check_trx_state(): Moved to a static function; partially replaced with
individual debug assertions implementing equivalent or stricter checks.
This is a backport of commit 7b51d11cca
from 10.5.
trx_t::will_lock: Changed the type to bool.
trx_t::is_autocommit_non_locking(): Replaces
trx_is_autocommit_non_locking().
trx_is_ac_nl_ro(): Remove (replaced with equivalent assertion expressions).
assert_trx_nonlocking_or_in_list(): Remove.
Replaced with at least as strict checks in each place.
check_trx_state(): Moved to a static function; partially replaced with
individual debug assertions implementing equivalent or stricter checks.
Between btr_pcur_store_position() and btr_pcur_restore_position()
it is possible that purge empties a table and enlarges
index->n_core_fields and index->n_core_null_bytes.
Therefore, we must cache index->n_core_fields in
btr_pcur_t::old_n_core_fields so that btr_pcur_t::old_rec can be
parsed correctly.
Unfortunately, this is a huge change, because we will replace
"bool leaf" parameters with "ulint n_core"
(passing index->n_core_fields, or 0 for non-leaf pages).
For special cases where we know that index->is_instant() cannot hold,
we may also pass index->n_fields.
Before MDEV-14638, there was no race condition between the
execution of fetch_data_into_cache() and transaction commit.
fetch_data_into_cache(): Acquire trx_t::mutex before checking
trx_t::state, to prevent a concurrent transition from
TRX_STATE_COMMITTED_IN_MEMORY to TRX_STATE_NOT_STARTED
in trx_commit_in_memory().
A performance regression was introduced by
commit e71e613353 (MDEV-24671)
and mostly addressed by
commit 455514c800.
The regression is likely caused by increased contention
lock_sys.latch (former lock_sys.mutex), possibly indirectly
caused by contention on lock_sys.wait_mutex. This change aims to
reduce both, but further improvements will be needed.
lock_wait(): Minimize the lock_sys.wait_mutex hold time.
lock_sys_t::deadlock_check(): Add a parameter for indicating
whether lock_sys.latch is exclusively locked.
trx_t::was_chosen_as_deadlock_victim: Always use atomics.
lock_wait_wsrep(): Assume that no mutex is being held.
Deadlock::report(): Always kill the victim transaction.
lock_sys_t::timeout: New counter to back MONITOR_TIMEOUT.
For now, we will acquire the lock_sys.latch only in exclusive mode,
that is, use it as a mutex.
This is preparation for the next commit where we will introduce
a less intrusive alternative, combining a shared lock_sys.latch
with dict_table_t::lock_mutex or a mutex embedded in
lock_sys.rec_hash, lock_sys.prdt_hash, or lock_sys.prdt_page_hash.
lock_wait(): Replaces lock_wait_suspend_thread(). Wait for the lock to
be granted or the transaction to be killed using mysql_cond_timedwait()
or mysql_cond_wait().
lock_wait_end(): Replaces que_thr_end_lock_wait() and
lock_wait_release_thread_if_suspended().
lock_wait_timeout_task: Remove. The operating system kernel will
resume the mysql_cond_timedwait() in lock_wait(). An added benefit
is that innodb_lock_wait_timeout no longer has a 'jitter' of 1 second,
which was caused by this wake-up task waking up only once per second,
and then waking up any threads for which the timeout (which was only
measured in seconds) was exceeded.
innobase_kill_query(): Set trx->error_state=DB_INTERRUPTED,
so that a call trx_is_interrupted(trx) in lock_wait() can be avoided.
We will protect things more consistently with lock_sys.wait_mutex,
which will be moved below lock_sys.mutex in the latching order.
trx_lock_t::cond: Condition variable for !wait_lock, used with
lock_sys.wait_mutex.
srv_slot_t: Remove. Replaced by trx_lock_t::cond,
lock_grant_after_reset(): Merged to to lock_grant().
lock_rec_get_index_name(): Remove.
lock_sys_t: Introduce wait_pending, wait_count, wait_time, wait_time_max
that are protected by wait_mutex.
trx_lock_t::que_state: Remove.
que_thr_state_t: Remove QUE_THR_COMMAND_WAIT, QUE_THR_LOCK_WAIT.
que_thr_t: Remove is_active, start_running(), stop_no_error().
que_fork_t::n_active_thrs, trx_lock_t::n_active_thrs: Remove.
Any transaction that has requested a lock must have trx->id!=0.
trx_print_low(): Distinguish non-locking or inactive transaction
objects by displaying the pointer in parentheses.
fill_trx_row(): Do not try to map trx->id to a pointer-based value.
SHOW ENGINE INNODB MUTEX functionality is completely removed,
as are the InnoDB latching order checks.
We will enforce innodb_fatal_semaphore_wait_threshold
only for dict_sys.mutex and lock_sys.mutex.
dict_sys_t::mutex_lock(): A single entry point for dict_sys.mutex.
lock_sys_t::mutex_lock(): A single entry point for lock_sys.mutex.
FIXME: srv_sys should be removed altogether; it is duplicating tpool
functionality.
fil_crypt_threads_init(): To prevent SAFE_MUTEX warnings, we must
not hold fil_system.mutex.
fil_close_all_files(): To prevent SAFE_MUTEX warnings for
fil_space_destroy_crypt_data(), we must not hold fil_system.mutex
while invoking fil_space_free_low() on a detached tablespace.
Let us replace os_event_t with mysql_cond_t, and replace the
necessary ib_mutex_t with mysql_mutex_t so that they can be
used with condition variables.
Also, let us replace polling (os_thread_sleep() or timed waits)
with plain mysql_cond_wait() wherever possible.
Furthermore, we will use the lightweight srw_mutex for trx_t::mutex,
to hopefully reduce contention on lock_sys.mutex.
FIXME: Add test coverage of
mariabackup --backup --kill-long-queries-timeout
InnoDB buffer pool block and index tree latches depend on a
special kind of read-update-write lock that allows reentrant
(recursive) acquisition of the 'update' and 'write' locks
as well as an upgrade from 'update' lock to 'write' lock.
The 'update' lock allows any number of reader locks from
other threads, but no concurrent 'update' or 'write' lock.
If there were no requirement to support an upgrade from 'update'
to 'write', we could compose the lock out of two srw_lock
(implemented as any type of native rw-lock, such as SRWLOCK on
Microsoft Windows). Removing this requirement is very difficult,
so in commit f7e7f487d4b06695f91f6fbeb0396b9d87fc7bbf we
implemented an 'update' mode to our srw_lock.
Re-entrant or recursive locking is mostly needed when writing or
freeing BLOB pages, but also in crash recovery or when merging
buffered changes to an index page. The re-entrancy allows us to
attach a previously acquired page to a sub-mini-transaction that
will be committed before whatever else is holding the page latch.
The SUX lock supports Shared ('read'), Update, and eXclusive ('write')
locking modes. The S latches are not re-entrant, but a single S latch
may be acquired even if the thread already holds an U latch.
The idea of the U latch is to allow a write of something that concurrent
readers do not care about (such as the contents of BTR_SEG_LEAF,
BTR_SEG_TOP and other page allocation metadata structures, or
the MDEV-6076 PAGE_ROOT_AUTO_INC). (The PAGE_ROOT_AUTO_INC field
is only updated when a dict_table_t for the table exists, and only
read when a dict_table_t for the table is being added to dict_sys.)
block_lock::u_lock_try(bool for_io=true) is used in buf_flush_page()
to allow concurrent readers but no concurrent modifications while the
page is being written to the data file. That latch will be released
by buf_page_write_complete() in a different thread. Hence, we use
the special lock owner value FOR_IO.
The index_lock::u_lock() improves concurrency on operations that
involve non-leaf index pages.
The interface has been cleaned up a little. We will use
x_lock_recursive() instead of x_lock() when we know that a
lock is already held by the current thread. Similarly,
a lock upgrade from U to X is only allowed via u_x_upgrade()
or x_lock_upgraded() but not via x_lock().
We will disable the LatchDebug and sync_array interfaces to
InnoDB rw-locks.
The SEMAPHORES section of SHOW ENGINE INNODB STATUS output
will no longer include any information about InnoDB rw-locks,
only TTASEventMutex (cmake -DMUTEXTYPE=event) waits.
This will make a part of the 'innotop' script dead code.
The block_lock buf_block_t::lock will not be covered by any
PERFORMANCE_SCHEMA instrumentation.
SHOW ENGINE INNODB MUTEX and INFORMATION_SCHEMA.INNODB_MUTEXES
will no longer output source code file names or line numbers.
The dict_index_t::lock will be identified by index and table names,
which should be much more useful. PERFORMANCE_SCHEMA is lumping
information about all dict_index_t::lock together as
event_name='wait/synch/sxlock/innodb/index_tree_rw_lock'.
buf_page_free(): Remove the file,line parameters. The sux_lock will
not store such diagnostic information.
buf_block_dbg_add_level(): Define as empty macro, to be removed
in a subsequent commit.
Unless the build was configured with cmake -DPLUGIN_PERFSCHEMA=NO
the index_lock dict_index_t::lock will be instrumented via
PERFORMANCE_SCHEMA. Similar to
commit 1669c8890c
we will distinguish lock waits by registering shared_lock,exclusive_lock
events instead of try_shared_lock,try_exclusive_lock.
Actual 'try' operations will not be instrumented at all.
rw_lock_list: Remove. After MDEV-24167, this only covered
buf_block_t::lock and dict_index_t::lock. We will output their
information by traversing buf_pool or dict_sys.
Let us try to avoid code bloat for the common case that
performance_schema is disabled at runtime, and use
ATTRIBUTE_NOINLINE member functions for instrumented latch acquisition.
Also, let us distinguish lock waits from non-contended lock requests
by using write_lock,read_lock for the requests that lead to waits,
and try_write_lock,try_read_lock for the wait-free lock acquisitions.
Actual 'try' operations are not being instrumented at all.
We can simply use C++11 std::atomic for avoiding undefined behaviour
related to concurrent stores to a shared variable. On most if not all
ISAs, std::memory_order_relaxed loads and stores will not really
differ from non-atomic loads or stores.
Since commit 8ccb3caafb it should be
more efficient to use page_id_t rather than two separate variables
for tablespace identifier and page number.
lock_rec_fold(): Replaced with page_id_t::fold().
lock_rec_hash(): Replaced with lock_sys.hash(page_id).
lock_rec_expl_exist_on_page(), lock_rec_get_first_on_page_addr(),
lock_rec_get_first_on_page(): Replaced with lock_sys.get_first().
The parameters innodb_thread_concurrency and innodb_commit_concurrency
were useful years ago when both computing resources and the implementation
of some shared data structures were limited. MySQL 5.0 or 5.1 had trouble
scaling beyond 8 concurrent connections. Most of the scalability bottlenecks
have been removed since then, and the transactions per second delivered
by MariaDB Server 10.5 should not dramatically drop upon exceeding the
'optimal' number of connections.
Hence, enabling any concurrency throttling for InnoDB actually makes
things worse. We have seen many customers mistakenly setting this to a
small value like 16 or 64 and then complaining the server was slow.
Ignoring the parameters allows us to remove some normally unused code
and data structures, which could slightly improve performance.
innodb_thread_concurrency, innodb_commit_concurrency,
innodb_replication_delay, innodb_concurrency_tickets,
innodb_thread_sleep_delay, innodb_adaptive_max_sleep_delay:
Deprecate and ignore; hard-wire to 0.
The column INFORMATION_SCHEMA.INNODB_TRX.trx_concurrency_tickets
will always report 0.
commit d09aec7a15 (MDEV-19940)
caused a regression. We made wait_lock_get_heap_no() return
uint16_t instead of ulint, and we mostly replaced the previous
magic value ULINT_UNDEFINED with 0. But, we failed to adjust
some assertions. Furthermore, 0 is a valid although rare value
for record locks. (Record locks can be temporarily stored on
page infimum in some operations that involve multiple leaf pages.)
Let us use 0xFFFF as the magic value. Valid heap numbers
are limited to less than 9362 = innodb_page_size/(5+1+1)
when using a minimal 1-byte PRIMARY KEY and a
secondary index on a NULL or '' column.
This was the last abuse of trx_sys.mutex, which is now exclusively
protecting trx_sys.trx_list.
This global acquisition was also potential scalability bottleneck for
oltp_read_write benchmark. Although it didn't expose itself as such due
to larger scalability issues.
Replaced trx_sys.mutex based synchronisation between ReadView creator
thread and purge coordinator thread performing latest view clone with
ReadView::m_mutex.
It also allowed to simplify tri-state view m_state down to boolean
m_open flag.
For performance reasons trx->read_view.close() is left as atomic relaxed
store, so that we don't have to waste resources for otherwise meaningless
mutex acquisition.
innobase_get_charset(), innobase_get_stmt_safe(): Remove.
It is more efficient and readable to invoke thd_charset()
and thd_query_safe() directly, without a non-inlined wrapper function.
The -Wconversion in GCC seems to be stricter than in clang.
GCC at least since version 4.4.7 issues truncation warnings for
assignments to bitfields, while clang 10 appears to only issue
warnings when the sizes in bytes rounded to the nearest integer
powers of 2 are different.
Before GCC 10.0.0, -Wconversion required more casts and would not
allow some operations, such as x<<=1 or x+=1 on a data type that
is narrower than int.
GCC 5 (but not GCC 4, GCC 6, or any later version) is complaining
about x|=y even when x and y are compatible types that are narrower
than int. Hence, we must rewrite some x|=y as
x=static_cast<byte>(x|y) or similar, or we must disable -Wconversion.
In GCC 6 and later, the warning for assigning wider to bitfields
that are narrower than 8, 16, or 32 bits can be suppressed by
applying a bitwise & with the exact bitmask of the bitfield.
For older GCC, we must disable -Wconversion for GCC 4 or 5 in such
cases.
The bitwise negation operator appears to promote short integers
to a wider type, and hence we must add explicit truncation casts
around them. Microsoft Visual C does not allow a static_cast to
truncate a constant, such as static_cast<byte>(1) truncating int.
Hence, we will use the constructor-style cast byte(~1) for such cases.
This has been tested at least with GCC 4.8.5, 5.4.0, 7.4.0, 9.2.1, 10.0.0,
clang 9.0.1, 10.0.0, and MSVC 14.22.27905 (Microsoft Visual Studio 2019)
on 64-bit and 32-bit targets (IA-32, AMD64, POWER 8, POWER 9, ARMv8).
