- InnoDB mistakenly identifies the non-unique FTS_DOC_ID index as
FTS_DOC_ID_INDEX while loading the table. dict_load_indexes()
should check whether the index is unique before assigning
fts_doc_id_index
Before version 10, GCC would think that a right shift of an
unsigned char returns int. Let us explicitly cast that back,
to silence a bogus -Wconversion warning.
- In case of discarded tablespace, InnoDB can't read the root page to
assign the n_core_null_bytes. Consecutive instant DDL fails because
of non-matching n_core_null_bytes.
fil_space_t::acquire_low(): Introduce a parameter that specifies
which flags should be avoided. At all times, referenced() must not
be incremented if the STOPPING flag is set. When fil_system.mutex
is not being held by the current thread, the reference must not be
incremented if the CLOSING flag is set (unless NEEDS_FSYNC is set,
in fil_space_t::flush()).
fil_space_t::acquire(): Invoke acquire_low(STOPPING | CLOSING).
In this way, the reference count cannot be incremented after
fil_space_t::try_to_close() invoked fil_space_t::set_closing().
If the CLOSING flag was set, we must retry acquire_low() after
acquiring fil_system.mutex.
fil_space_t::prepare_acquired(): Replaces prepare(true).
fil_space_t::acquire_and_prepare(): Replaces prepare().
This basically retries fil_space_t::acquire() after
acquiring fil_system.mutex.
- Redundant InnoDB table fails to set the flags2 while loading
the table. It leads to "Upgrade index name failure" during alter
operation. InnoDB should set the flags2 to FTS_AUX_HEX_NAME when
fts is being loaded
recv_sys_t::recover_deferred(): Hold the exclusive page latch until
the tablespace has been set up. Otherwise, the write of the page
may be lost due to non-existent tablespace. This race only affects
the recovery of the first page in a newly created tablespace.
This race condition was introduced in MDEV-24626.
The Spider mixes the comma join with other join types, and thus
ERROR 1054 occurs. This is well-known issue caused by the higher
precedence of JOIN over the comma (,).
We can fix the problem simply by using JOINs instead of commas.
The bug is caused by a similar mechanism as MDEV-21027.
The function, check_insert_or_replace_autoincrement, failed to open
all the partitions on INSERT SELECT statements and it results in the
assertion error.
The heap-use-after-free is caused by the following mechanism:
* In the execution of FLUSH TABLE WITH READ LOCK, the function
spider_free_trx_conn() is called and the connections held by
SPIDER_TRX::trx_conn_hash are freed.
* Then, an instance of ha_spider maintains the freed connections
because they are also referenced from ha_spider::conns.
The ha_spider instance is kept in a lock structure until the
corresponding table is unlocked.
* Spider accesses ha_spider::conns on the implicit UNLOCK TABLE
issued by BEGIN.
In the first place, when the connections have been freed, it means
that there are really no remote table locked by Spider.
Thus, there is no need for Spider to access ha_spider::cons on the
implicit UNLOCK TABLE.
We can fix the bug by removing the above mentioned access to
ha_spider::conns. We also modified spider_free_trx_conn() so that it
frees the connections only when no table is locked to reduce the
chance of another heap-use-after-free on ha_spider::conns.
prepare_inplace_alter_table_dict(): If the table will not be rebuilt,
preserve all of the original ROW_FORMAT, including the compressed
page size flags related to ROW_FORMAT=COMPRESSED.
btr_root_raise_and_insert(), btr_lift_page_up(),
rtr_page_split_and_insert(): Reset DB_FAIL from a failure to
copy records on a ROW_FORMAT=COMPRESSED page to DB_SUCCESS
before retrying.
This fixes a regression that was introduced by
commit 0b47c126e3 (MDEV-13542).
btr_root_raise_and_insert(): Remove a redundant condition.
btr_page_split_and_insert() will invoke btr_page_split_and_insert()
if needed.
buf_page_print(): Dump the buffer page 32 bytes (64 hexadecimal digits)
per line. In this way, the limitation in mtr
("Data too long for column 'line'") will not be triggered.
Also, do not bother decoding the page contents, because everything
is present in the hexadecimal output.
dict_index_find_on_id_low(): Merge to dict_index_get_if_in_cache_low().
The direct call in buf_page_print() was prone to crashing, in case the
table definition was concurrently evicted or dropped from the
data dictionary cache.
Spider supports (or at least allows) INSERT DELAYED but the
documentation does not specify spider as a storage engine that supports
"INSERT DELAYED".
Also, although not mentioned in the documentation, "INSERT DELAYED" is
not intended to be executed inside a transaction, as can be seen from
the list of supported storage engines.
The current implementation allows executing a delayed insert on a
remote transactional table and this breaks the consistency ensured by
the transaction.
We too remove "internal_delayed", one of the Spider table parameters.
Documentation says,
> Whether to transmit existence of delay to remote servers when
> executing an INSERT DELAYED statement on local server.
This table parameter is only used for "INSERT DELAYED".
Reviewed by: Nayuta Yanagisawa
When the variable, spider_semi_table_lock, is 1, Spider sends
LOCK TABLES before each SQL execution. The feature is for
non-transactional remote tables and adds some overhead to query
executions.
We change the default value of the plugin variable to 0 and then
deprecate the variable because it is rare to use non-transactional
engines these days and the variable complicates the code.
The variable, spider_semi_table_lock_connection, should be too
deprecated because it is for tweaking the semi-table locking.
Deprecate the high availability feature of Spider as the feature is
beyond the scope of the federated storage engine and it complicates
the implementation of Spider.
Revert "TSAN: data race on vptr (ctor/dtor vs virtual call)"
This reverts commit 78084fa747.
This commit was done to please TSAN, which falsely reported an error
where there was none.
Yet as consequence, it could cause a real error, a crash in os_aio_free on
shutdown
fil_name_process(): If the recovery of a tablespace was deferred,
do invoke fil_ibd_load() even though the name in recv_spaces is
not changing. This allows us to recover from a situation where
there are many FILE_RENAME records, renaming a tablespace back
and forth, and a FILE_MODIFY record that had been written by
fil_names_clear().
Co-developed with: Thirunarayanan Balathandayuthapani
trx_undo_rec_copy(): Return nullptr if the undo record is corrupted.
trx_undo_rec_get_undo_no(): Define inline with the declaration.
trx_purge_dummy_rec: Replaced with a -1 pointer.
row_undo_rec_get(), UndorecApplier::apply_undo_rec(): Check
if trx_undo_rec_copy() returned nullptr.
trx_purge_get_next_rec(): Return nullptr upon encountering any
corruption, to signal the end of purge.
On GNU/Linux, even though the C11 aligned_alloc() appeared in
GNU libc early on, some custom memory allocators did not
implement it until recently. For example, before
gperftools/gperftools@d406f22853
the free() in tcmalloc would fail to free memory that was
returned by aligned_alloc(), because the latter would map to the
built-in allocator of libc. The Linux specific memalign() has a
similar interface and is safer to use, because it has been
available for a longer time. For AddressSanitizer, we will use
aligned_alloc() so that the constraint on size can be enforced.
buf_tmp_reserve_compression_buf(): When HAVE_ALIGNED_ALLOC holds,
round up the size to be an integer multiple of the alignment.
pfs_malloc(): In the unit test stub, round up the size to be an
integer multiple of the alignment.
Table_cache_instance: Define the structure aligned at
the CPU cache line, and remove a pad[] data member.
Krunal Bauskar reported this to improve performance on ARMv8.
aligned_malloc(): Wrapper for the Microsoft _aligned_malloc()
and the ISO/IEC 9899:2011 <stdlib.h> aligned_alloc().
Note: The parameters are in the Microsoft order (size, alignment),
opposite of aligned_alloc(alignment, size).
Note: The standard defines that size must be an integer multiple
of alignment. It is enforced by AddressSanitizer but not by GNU libc
on Linux.
aligned_free(): Wrapper for the Microsoft _aligned_free() and
the standard free().
HAVE_ALIGNED_ALLOC: A new test. Unfortunately, support for
aligned_alloc() may still be missing on some platforms.
We will fall back to posix_memalign() for those cases.
HAVE_MEMALIGN: Remove, along with any use of the nonstandard memalign().
PFS_ALIGNEMENT (sic): Removed; we will use CPU_LEVEL1_DCACHE_LINESIZE.
PFS_ALIGNED: Defined using the C++11 keyword alignas.
buf_pool_t::page_hash_table::create(),
lock_sys_t::hash_table::create():
lock_sys_t::hash_table::resize(): Pad the allocation size to an
integer multiple of the alignment.
Reviewed by: Vladislav Vaintroub
There was a race condition between log_checkpoint_low() and
deleting or renaming data files. The scenario is as follows:
1. The buffer pool does not contain dirty pages.
2. A FILE_DELETE or FILE_RENAME record is written.
3. The checkpoint LSN will be moved ahead of the write of the record.
4. The server is killed before the file is actually renamed or deleted.
We will prevent this race condition by ensuring that a log checkpoint
cannot occur between the durable write and the file system operation:
1. Durably write the FILE_DELETE or FILE_RENAME record.
2. Perform the file system operation.
3. Allow any log checkpoint to proceed.
mtr_t::commit_file(): Implement the DELETE or RENAME logic.
fil_delete_tablespace(): Delegate some of the logic to
mtr_t::commit_file().
fil_space_t::rename(): Delegate some logic to mtr_t::commit_file().
Remove the debug injection point fil_rename_tablespace_failure_2
because we do test RENAME failures without any debug injection.
fil_name_write_rename_low(), fil_name_write_rename(): Remove.
Tested by Matthias Leich
ha_innobase::prepare_inplace_alter_table(): If the ALTER TABLE
operation is no-op for InnoDB, do reset m_prebuilt->trx_id
so that ha_innobase::table_version() will always report either 0
or the identifier of the transaction that would commit changes
to the InnoDB data dictionary.
The failure scenario involved a completed DROP INDEX followed by
a no-op ALTER TABLE during which the server was killed.
An effort to create a reproducible test failed.
ha_partition::set_auto_increment_if_higher expects
part_share->auto_inc_initialized is true or can_use_for_auto_inc_init()
is false (but as the comment of this method says, it returns false
only if we use Spider engine with DROP TABLE or ALTER TABLE query).
However, part_share->auto_inc_initialized becomes true only after all
partitions are opened (since 6dce6aeceb).
Therefore, I added a conditional expression in order to read all
partitions when we execute REPLACE on a table that has an
AUTO_INCREMENT column.
Reviewed by: Nayuta Yanagisawa
Reviewed by: Alexey Botchkov
Since commit 0b47c126e3 (MDEV-13542)
we treat all-zero pages as corrupted ones.
During a stress test, a read-ahead of an all-zero page was triggered
and the page read was completed concurrently with buf_page_create_low().
This caused the assertion to fail, because buf_page_create_low() was
waiting for the page latch.
buf_page_get_low(): Only invoke buf_pool_t::corrupted_evict()
if the block was not already marked as corrupted.
buf_page_create_low(): On page identifier mismatch, retry the
buf_pool.page_hash lookup.
buf_pool_t::corrupted_evict(): Set the state of the block to FREED
so that a concurrent buf_page_get_low() will refuse to load the page.
Wait for the page latch to be vacant before proceeding to remove
the block from buf_pool.page_hash and buf_pool.LRU.
page_id_t::set_corrupted(), page_id_t::is_corrupted(): Accessors
for indicating a corrupted page identifier.
Tested by Matthias Leich