when opening 10.1- table that has virtual columns:
1. don't error out if it has vcols over autoinc columns.
just issue a warning.
2. set vcol type properly
3. in innodb: use table->s->stored_fields instead of table->s->fields,
because that's what was stored in innodb data dictionary
don't use thd->query_id check in background purge threads
(it doesn't work, because thd->query_id is never incremented there)
instead use thd->open_tables directly, there can be only one table
there anyway, and this is the table opened by this purge thread.
When using innodb_page_size=16k, InnoDB tables
that were created in MariaDB 10.1.0 to 10.1.20 with
PAGE_COMPRESSED=1 and
PAGE_COMPRESSION_LEVEL=2 or PAGE_COMPRESSION_LEVEL=3
would fail to load.
fsp_flags_is_valid(): When using innodb_page_size=16k, use a
more strict check for .ibd files, with the assumption that
nobody would try to use different-page-size files.
When using innodb_page_size=16k, InnoDB tables
that were created in MariaDB 10.1.0 to 10.1.20 with
PAGE_COMPRESSED=1 and
PAGE_COMPRESSION_LEVEL=2 or PAGE_COMPRESSION_LEVEL=3
would fail to load.
fsp_flags_is_valid(): When using innodb_page_size=16k, use a
more strict check for .ibd files, with the assumption that
nobody would try to use different-page-size files.
This is a regression caused by
commit bb60a832ed
srv_shutdown_all_bg_threads(): If os_thread_count indicates that
no threads are running, do not bother checking thread status.
This avoids a crash when InnoDB startup is aborted before
os_aio_init() has been invoked. (os_aio_all_slots_free() would
dereference AIO::s_reads even though it is NULL.)
InnoDB I/O and buffer pool interfaces and the redo log format
have been changed between MariaDB 10.1 and 10.2, and the backup
code has to be adjusted accordingly.
The code has been simplified, and many memory leaks have been fixed.
Instead of the file name xtrabackup_logfile, the file name ib_logfile0
is being used for the copy of the redo log. Unnecessary InnoDB startup and
shutdown and some unnecessary threads have been removed.
Some help was provided by Vladislav Vaintroub.
Parameters have been cleaned up and aligned with those of MariaDB 10.2.
The --dbug option has been added, so that in debug builds,
--dbug=d,ib_log can be specified to enable diagnostic messages
for processing redo log entries.
By default, innodb_doublewrite=OFF, so that --prepare works faster.
If more crash-safety for --prepare is needed, double buffering
can be enabled.
The parameter innodb_log_checksums=OFF can be used to ignore redo log
checksums in --backup.
Some messages have been cleaned up.
Unless --export is specified, Mariabackup will not deal with undo log.
The InnoDB mini-transaction redo log is not only about user-level
transactions; it is actually about mini-transactions. To avoid confusion,
call it the redo log, not transaction log.
We disable any undo log processing in --prepare.
Because MariaDB 10.2 supports indexed virtual columns, the
undo log processing would need to be able to evaluate virtual column
expressions. To reduce the amount of code dependencies, we will not
process any undo log in prepare.
This means that the --export option must be disabled for now.
This also means that the following options are redundant
and have been removed:
xtrabackup --apply-log-only
innobackupex --redo-only
In addition to disabling any undo log processing, we will disable any
further changes to data pages during --prepare, including the change
buffer merge. This means that restoring incremental backups should
reliably work even when change buffering is being used on the server.
Because of this, preparing a backup will not generate any further
redo log, and the redo log file can be safely deleted. (If the
--export option is enabled in the future, it must generate redo log
when processing undo logs and buffered changes.)
In --prepare, we cannot easily know if a partial backup was used,
especially when restoring a series of incremental backups. So, we
simply warn about any missing files, and ignore the redo log for them.
FIXME: Enable the --export option.
FIXME: Improve the handling of the MLOG_INDEX_LOAD record, and write
a test that initiates a backup while an ALGORITHM=INPLACE operation
is creating indexes or rebuilding a table. An error should be detected
when preparing the backup.
FIXME: In --incremental --prepare, xtrabackup_apply_delta() should
ensure that if FSP_SIZE is modified, the file size will be adjusted
accordingly.
The POINT data type is being treated just like any other
geometry data type in InnoDB. The fixed-length data type
DATA_POINT had been introduced in WL#6942 based on a
misunderstanding and without appropriate review.
Because of fundamental design problems (such as a
DEFAULT POINT(0 0) value secretly introduced by InnoDB),
the code was disabled in Oracle Bug#20415831 fix.
This patch removes the dead code and definitions that were
left behind by the Oracle Bug#20415831 patch.
This is preparation for MDEV-12288, which would set DB_TRX_ID=0
when purging history. Also with that change in place, delete-marked
records must always refer to an undo log record via a nonzero
DB_TRX_ID column. (The DB_TRX_ID is only present in clustered index
leaf page records.)
btr_cur_parse_del_mark_set_clust_rec(), rec_get_trx_id():
Statically allocate the offsets
(should never use the heap). Add some debug assertions.
Replace some use of rec_get_trx_id() with row_get_rec_trx_id().
trx_undo_report_row_operation(): Add some sanity checks that are
common for all operations that produce undo log.
The field fts_token->position is not initialized in
row_merge_fts_doc_tokenize(). We cannot have that field
without changing the fulltext parser plugin ABI
(adding st_mysql_ftparser_boolean_info::position,
as it was done in MySQL 5.7 in WL#6943).
The InnoDB fulltext parser plugins "ngram" and "Mecab" that were
introduced in MySQL 5.7 do depend on that field. But the simple_parser
does not. Apparently, simple_parser is leaving the field as 0.
So, in our fix we will assume that the missing position field is 0.
In Mariabackup, we would want the backed-up redo log file size to be
a multiple of 512 bytes, or OS_FILE_LOG_BLOCK_SIZE. However, at startup,
InnoDB would be picky, requiring the file size to be a multiple of
innodb_page_size.
Furthermore, InnoDB would require the parameter to be a multiple of
one megabyte, while the minimum granularity is 512 bytes. Because
the data-file-oriented fil_io() API is being used for writing the
InnoDB redo log, writes will for now require innodb_log_file_size to
be a multiple of the maximum innodb_page_size (65536 bytes).
To complicate matters, InnoDB startup divided srv_log_file_size by
UNIV_PAGE_SIZE, so that initially, the unit was bytes, and later it
was innodb_page_size. We will simplify this and keep srv_log_file_size
in bytes at all times.
innobase_log_file_size: Remove. Remove some obsolete checks against
overflow on 32-bit systems. srv_log_file_size is always 64 bits, and
the maximum size 512GiB in multiples of innodb_page_size always fits
in ulint (which is 32 or 64 bits). 512GiB would be 8,388,608*64KiB or
134,217,728*4KiB.
log_init(): Remove the parameter file_size that was always passed as
srv_log_file_size.
log_set_capacity(): Add a parameter for passing the requested file size.
srv_log_file_size_requested: Declare static in srv0start.cc.
create_log_file(), create_log_files(),
innobase_start_or_create_for_mysql(): Invoke fil_node_create()
with srv_log_file_size expressed in multiples of innodb_page_size.
innobase_start_or_create_for_mysql(): Require the redo log file sizes
to be multiples of 512 bytes.
trx_sys_print_mysql_binlog_offset(): Use 64-bit arithmetics and ib::info().
TRX_SYS_MYSQL_LOG_OFFSET: Replaces TRX_SYS_MYSQL_LOG_OFFSET_HIGH,
TRX_SYS_MYSQL_LOG_OFFSET_LOW.
trx_sys_update_mysql_binlog_offset(): Remove the constant parameter
field=TRX_SYS_MYSQL_LOG_INFO. Use 64-bit arithmetics.
A similar change was contributed to Percona XtraBackup, but for some
reason, it is not present in Percona XtraDB. Since MDEV-9566
(MariaDB 10.1.23), that change is present in the MariaDB XtraDB.
recv_sys_init(): Remove the parameter.
recv_sys_create(): Merge to recv_sys_init().
recv_sys_mem_free(): Merge to recv_sys_close().
log_mem_free(): Merge to log_shutdown().
srv_log_files_created: A debug flag to ensure that InnoDB redo log
files can only be created once in the server lifetime, and that
after log files have been created, no crash recovery will take place.
recv_scan_log_recs(): Detect the special case where the log consists
of a sole MLOG_CHECKPOINT record, such as immediately after creating
the redo logs.
recv_recovery_from_checkpoint_start(): Skip the recovery message
if the redo log is logically empty.
A merge error caused InnoDB bootstrap to fail when
innodb_undo_tablespaces was set to more than 2.
This was because of a bug that was introduced to
srv_undo_tablespaces_init() by the merge.
Furthermore, some adjustments for Oracle Bug#25551311 aka
Bug#23517560 changes were forgotten. We must minimize direct
references to srv_undo_tablespaces_open and use predicates
instead.
srv_undo_tablespaces_init(): Increment srv_undo_tablespaces_open
once, not twice, per loop iteration.
is_system_or_undo_tablespace(): Remove (unused function).
is_predefined_tablespace(): Invoke srv_is_undo_tablespace().
When it comes to DEFAULT values of columns, InnoDB is imposing both
unnecessary and insufficient conditions on whether ALGORITHM=INPLACE
should be allowed for ALTER TABLE.
When changing an existing column to NOT NULL, any NULL values in the
columns only get a special treatment if the column is changed to an
AUTO_INCREMENT column (which is not supported by ALGORITHM=INPLACE)
or the column type is TIMESTAMP. In all other cases, an error
must be reported for the failure to convert a NULL value to NOT NULL.
InnoDB was unnecessarily interested in whether the DEFAULT value
is not constant when altering other than TIMESTAMP columns. Also,
when changing a TIMESTAMP column to NOT NULL, InnoDB was performing
an insufficient check, and it was incorrectly allowing a constant
DEFAULT value while not being able to replace NULL values with that
constant value.
Furthermore, in ADD COLUMN, InnoDB is unnecessarily rejecting certain
nondeterministic DEFAULT expressions (depending on the session
parameters or the current time).
buf_flush_page_cleaner_coordinator(): Signal the thread creator
that the error log output regarding setpriority() has been issued.
innobase_start_or_create_for_mysql(): Wait for
buf_flush_page_cleaner_coordinator() to completely start up.
This prevents sporadic failures of tests that search the server
error log for InnoDB redo log recovery messages.
While the primary purpose of innodb_force_recovery is to allow
data to be rescued from an InnoDB instance that would crash due
to some data corruption, the settings 1, 2, or 3 are relatively
safe to use and there is no need to prevent write transactions
in these modes.
The setting innodb_force_recovery=4 and above can cause database
corruption. For those modes, we already set the flag
high_level_read_only to disable modifications, except DROP TABLE.
MODIFICATIONS_NOT_ALLOWED_MSG_FORCE_RECOVERY: Remove. There is no
need to spam the error log for each refused DML operation. It suffices
to return an error to the client. There will be messages at startup
if innodb_read_only or innodb_force_recovery are preventing writes.
The original intention of the setting innodb_force_recovery=3 was to
disable background activity that could create trouble, most notably,
the rollback of incomplete transactions, and the purge of transaction
history.
MySQL 5.6 introduced more background threads, it is creating
dict_stats_thread and fts_optimize_thread even though these threads
are at least as non-essential as the rollback and purge. These
threads are in fact worse, because they can create new transactions
on their own.
innobase_start_or_create_for_mysql(): Do not create any internal
undo log sources unless innodb_force_recovery<3.
buf_flush_init_for_writing(): Reset the FIL_PAGE_TYPE
of the TRX_SYS page to the canonical value
FIL_PAGE_TYPE_TRX_SYS instead of FIL_PAGE_TYPE_UNKNOWN.
Comment from Codership:-
To fix the problem, we changed the certification logic in galera to treat insert
on child table row as exclusive to prevent any operation on referenced
parent table row. At the same time, update and delete on
child table row were demoted to "shared", which makes it possible to
update/delete referenced parent table row, but only in a later transaction.
This change allows somewhat more concurrency for foreign key constrained
transactions, but is still safe for correct certification end result.
log_calc_max_ages(): Use the requested size in the check, instead of
the detected redo log size. The redo log will be resized at startup
if it differs from what has been requested.
log_calc_max_ages(): Use the requested size in the check, instead of
the detected redo log size. The redo log will be resized at startup
if it differs from what has been requested.
When the btr_search_latch was split into an array of latches
in MySQL 5.7.8 as part of the Oracle Bug#20985298 fix, the "caching"
of the latch across storage engine API calls was removed, and
the field trx->has_search_latch would only be set during a short
time frame in the execution of row_search_mvcc(), which was
formerly called row_search_for_mysql().
This means that the column
INFORMATION_SCHEMA.INNODB_TRX.TRX_ADAPTIVE_HASH_LATCHED will always
report 0. That column cannot be removed in MariaDB 10.2, but it
can be removed in future releases.
trx_t::has_search_latch: Remove.
trx_assert_no_search_latch(): Remove.
row_sel_try_search_shortcut_for_mysql(): Remove a redundant condition
on trx->has_search_latch (it was always true).
sync_check_iterate(): Make the parameter const.
sync_check_functor_t: Make the operator() const, and remove result()
and the virtual destructor. There is no need to have mutable state
in the functors.
sync_checker<bool>: Replaces dict_sync_check and btrsea_sync_check.
sync_check: Replaces btrsea_sync_check.
dict_sync_check: Instantiated from sync_checker.
sync_allowed_latches: Use std::find() directly on the array.
Remove the std::vector.
TrxInInnoDB::enter(), TrxInInnoDB::exit(): Remove obviously redundant
debug assertions on trx->in_depth, and use equality comparison against 0
because it could be more efficient on some architectures.
The sole purpose of handlerton::release_temporary_latches and its wrapper
function was to release the InnoDB adaptive hash index latch
(btr_search_latch).
When the btr_search_latch was split into an array of latches
in MySQL 5.7.8 as part of the Oracle Bug#20985298 fix, the "caching"
of the latch across storage engine API calls was removed. As part of that,
the function trx_search_latch_release_if_reserved() was changed to an
assertion and the function trx_reserve_search_latch_if_not_reserved()
was removed, and handlerton::release_temporary_latches() practically
became a no-op.
Note: MDEV-12121 replaced the function
trx_search_latch_release_if_reserved()
with the more appropriately named macro trx_assert_no_search_latch().
dict_sys_tables_type_to_tf(): Change the parameter n_cols to not_redundant.
dict_tf_is_valid_not_redundant(): Refactored from dict_tf_is_valid().
dict_sys_tables_type_valid(): Replaces dict_sys_tables_type_validate().
Use the common function dict_tf_is_valid_not_redundant(), which validates
PAGE_COMPRESSION_LEVEL more strictly.
DICT_TF_GET_UNUSED(flags): Remove.
innodb.table_flags: Adjust the test case. Due to the MDEV-12873 fix
in 10.2, the corrupted flags for table test.td would be converted,
and a tablespace flag mismatch will occur when trying to open the file.
Remove the SHARED_SPACE flag that was erroneously introduced in
MariaDB 10.2.2, and shift the SYS_TABLES.TYPE flags back to where
they were before MariaDB 10.2.2. While doing this, ensure that
tables created with affected MariaDB versions can be loaded,
and also ensure that tables created with MySQL 5.7 using the
TABLESPACE attribute cannot be loaded.
MariaDB 10.2.2 picked the SHARED_SPACE flag from MySQL 5.7,
shifting the MariaDB 10.1 flags PAGE_COMPRESSION, PAGE_COMPRESSION_LEVEL,
ATOMIC_WRITES by one bit. The SHARED_SPACE flag would always
be written as 0 by MariaDB, because MariaDB does not support
CREATE TABLESPACE or CREATE TABLE...TABLESPACE for InnoDB.
So, instead of the bits AALLLLCxxxxxxx we would have
AALLLLC0xxxxxxx if the table was created with MariaDB 10.2.2
to 10.2.6. (AA=ATOMIC_WRITES, LLLL=PAGE_COMPRESSION_LEVEL,
C=PAGE_COMPRESSED, xxxxxxx=7 bits that were not moved.)
PAGE_COMPRESSED=NO implies LLLLC=00000. That is not a problem.
If someone created a table in MariaDB 10.2.2 or 10.2.3 with
the attribute ATOMIC_WRITES=OFF (value 2; AA=10) and without
PAGE_COMPRESSED=YES or PAGE_COMPRESSION_LEVEL, the table should be
rejected. We ignore this problem, because it should be unlikely
for anyone to specify ATOMIC_WRITES=OFF, and because 10.2.2 and
10.2.2 were not mature releases. The value ATOMIC_WRITES=ON (1)
would be interpreted as ATOMIC_WRITES=OFF, but starting with
MariaDB 10.2.4 the ATOMIC_WRITES attribute is ignored.
PAGE_COMPRESSED=YES implies that PAGE_COMPRESSION_LEVEL be between
1 and 9 and that ROW_FORMAT be COMPACT or DYNAMIC. Thus, the affected
wrong bit pattern in SYS_TABLES.TYPE is of the form AALLLL10DB00001
where D signals the presence of a DATA DIRECTORY attribute and B is 1
for ROW_FORMAT=DYNAMIC and 0 for ROW_FORMAT=COMPACT. We must interpret
this bit pattern as AALLLL1DB00001 (discarding the extraneous 0 bit).
dict_sys_tables_rec_read(): Adjust the affected bit pattern when
reading the SYS_TABLES.TYPE column. In case of invalid flags,
report both SYS_TABLES.TYPE (after possible adjustment) and
SYS_TABLES.MIX_LEN.
dict_load_table_one(): Replace an unreachable condition on
!dict_tf2_is_valid() with a debug assertion. The flags will already
have been validated by dict_sys_tables_rec_read(); if that validation
fails, dict_load_table_low() will have failed.
fil_ibd_create(): Shorten an error message about a file pre-existing.
Datafile::validate_to_dd(): Clarify an error message about tablespace
flags mismatch.
ha_innobase::open(): Remove an unnecessary warning message.
dict_tf_is_valid(): Simplify and stricten the logic. Validate the
values of PAGE_COMPRESSION. Remove error log output; let the callers
handle that.
DICT_TF_BITS: Remove ATOMIC_WRITES, PAGE_ENCRYPTION, PAGE_ENCRYPTION_KEY.
The ATOMIC_WRITES is ignored once the SYS_TABLES.TYPE has been validated;
there is no need to store it in dict_table_t::flags. The PAGE_ENCRYPTION
and PAGE_ENCRYPTION_KEY are unused since MariaDB 10.1.4 (the GA release
was 10.1.8).
DICT_TF_BIT_MASK: Remove (unused).
FSP_FLAGS_MEM_ATOMIC_WRITES: Remove (the flags are never read).
row_import_read_v1(): Display an error if dict_tf_is_valid() fails.
dict_table_t::thd: Remove. This was only used by btr_root_block_get()
for reporting decryption failures, and it was only assigned by
ha_innobase::open(), and never cleared. This could mean that if a
connection is closed, the pointer would become stale, and the server
could crash while trying to report the error. It could also mean
that an error is being reported to the wrong client. It is better
to use current_thd in this case, even though it could mean that if
the code is invoked from an InnoDB background operation, there would
be no connection to which to send the error message.
Remove dict_table_t::crypt_data and dict_table_t::page_0_read.
These fields were never read.
fil_open_single_table_tablespace(): Remove the parameter "table".
innodb.row_format_redundant: Really corrupt the SYS_TABLES.MIX_LEN,
and do not use any debug instrumentation. For tables created in the
system tablespace, the contents of the column will be ignored.
Only the table t1 will refuse to load.
dict_load_table_one(): Remove the DBUG_EXECUTE_IF instrumentation.
Omit a redundant error message "incorrect flags in SYS_TABLES".
dict_sys_tables_rec_read(): Partially revert the Oracle Bug#21644827
fix, and always report errors by the return value.
fts_create_in_mem_aux_table(): Do not rely on dict_table_t::flags2,
but instead evaluate the tablespace ID.
DICT_TF2_BITS: Reduce to the correct value of 7. The two extra
high-order bits were specific to MySQL 5.7.
Cover innodb.table_flags with the new innodb_page_size.combinations
32k and 64k.
dict_sys_tables_type_validate(): Remove an assertion that made a
check in the function redundant. Remove the excessive output to
the error log, as the invalid SYS_TABLES.TYPE value is already being
output.
