a large memory buffer on Windows
fil_extend_space_to_desired_size(), os_file_set_size(): Use calloc()
for memory allocation, and handle failures. Properly check the return
status of posix_fallocate().
On Windows, instead of extending the file by at most 1 megabyte at a time,
write a zero-filled page at the end of the file.
According to the Microsoft blog post
https://blogs.msdn.microsoft.com/oldnewthing/20110922-00/?p=9573
this will physically extend the file by writing zero bytes.
(InnoDB never uses DeviceIoControl() to set the file sparse.)
For innodb_plugin, port the XtraDB fix for MySQL Bug#56433
(introducing fil_system->file_extend_mutex). The bug was
fixed differently in MySQL 5.6 (and MariaDB Server 10.0).
Fix memory barrier issues on releasing mutexes. We must have a full
memory barrier between releasing a mutex lock and reading its waiters.
This prevents us from missing to release waiters due to reading the
number of waiters speculatively before releasing the lock. If threads
try and wait between us reading the waiters count and releasing the
lock, those threads might stall indefinitely.
Also, we must use proper ACQUIRE/RELEASE semantics for atomic
operations, not ACQUIRE/ACQUIRE.
Analysis: row_drop_table_for_mysql did not allow dropping
referenced table even in case when actual creating of the
referenced table was not successfull if foreign_key_checks=1.
Fix: Allow dropping referenced table even if foreign_key_checks=1
if actual table create returned error.
The root cause is that x86 has a stronger memory model than the ARM
processors. And the GCC builtins didn't issue the correct fences when
setting/unsetting the lock word. In particular during the mutex release.
The solution is rewriting atomic TAS operations: replace '__sync_' by
'__atomic_' if possible.
Reviewed-by: Sunny Bains <sunny.bains@oracle.com>
Reviewed-by: Bin Su <bin.x.su@oracle.com>
Reviewed-by: Debarun Banerjee <debarun.banerjee@oracle.com>
Reviewed-by: Krunal Bauskar <krunal.bauskar@oracle.com>
RB: 9782
RB: 9665
RB: 9783
There is several different ways to incorrectly define
foreign key constraint. In many cases earlier MariaDB
versions the error messages produced by these cases
are not very clear and helpful. This patch improves
the warning messages produced by foreign key parsing.
This is an addendum to the fix for MDEV-7026. The ARM memory model is
similar to that of PowerPC and thus needs the same semantics with
respect to memory barriers. That is, os_atomic_test_and_set_*_release()
must be a store with a release barrier followed by a full
barrier. Unlike x86 using __sync_lock_test_and_set() which is
implemented as “exclusive load with acquire barriers + exclusive store”
is insufficient in contexts where os_atomic_test_and_set_*_release()
macros are used.
Scenario:
1. The purge thread takes an undo log record and parses it and forms
the record to be purged. We have the primary and secondary keys
to locate the actual records.
2. Using the secondary index key, we search in the secondary index.
One record is found.
3. Then it is checked if this record can be purged. The answer is we
can purge this record. To determine this we look up the clustered
index record. Either there is no corresponding clustered index
record, or the matching clustered index record is delete marked.
4. Then we check whether the secondary index record is delete marked.
We find that it is not delete marked. We report warning in optimized
build and assert in debug build.
Problem:
In step 3, we report that the record is purgeable even though it is
not delete marked. This is because of inconsistency between the
following members of purge_node_t structure - found_clust, ref and pcur.
Solution:
In the row_purge_reposition_pcur(), if the persistent cursor restore
fails, then reset the purge_node_t->found_clust member. This will
keep the members of purge_node_t structure in a consistent state.
rb#8813 approved by Marko.
On PPC64 high-loaded server may crash due to assertion failure in InnoDB
rwlocks code.
This happened because load order between "recursive" and "writer_thread"
wasn't properly enforced.
innodb_stats_sample_pages
Analysis: If you set the number of analyzed pages
to very low number compared to actual pages on
that table/index it randomly pics those pages
(default 8 pages), this leads to fact that query
after analyze table returns different results. If
the index tree is small, smaller than 10 *
n_sample_pages + total_external_size, then the
estimate is ok. For bigger index trees it is
common that we do not see any borders between
key values in the few pages we pick. But still
there may be n_sample_pages different key values,
or even more. And it just tries to
approximate to n_sample_pages (8).
Fix: (1) Introduced new dynamic configuration variable
innodb_stats_sample_traditional that retains
the current design. Default false.
(2) If traditional sample is not used we use
n_sample_pages = max(min(srv_stats_sample_pages,
index->stat_index_size),
log2(index->stat_index_size)*
srv_stats_sample_pages);
(3) Introduced new dynamic configuration variable
stat_modified_counter (default = 0) if set
sets lower bound for row updates when statistics is re-estimated.
If user has provided upper bound for how many rows needs to be updated
before we calculate new statistics we use minimum of provided value
and 1/16 of table every 16th round. If no upper bound is provided
(srv_stats_modified_counter = 0, default) then calculate new statistics
if 1 / 16 of table has been modified
since the last time a statistics batch was run.
We calculate statistics at most every 16th round, since we may have
a counter table which is very small and updated very often.
@param t table
@return true if the table has changed too much and stats need to be
recalculated
*/
#define DICT_TABLE_CHANGED_TOO_MUCH(t) \
((ib_int64_t) (t)->stat_modified_counter > (srv_stats_modified_counter ? \
ut_min(srv_stats_modified_counter, (16 + (t)->stat_n_rows / 16)) : \
16 + (t)->stat_n_rows / 16))
The bug was that full memory barrier was missing in the code that ensures that
a waiter on an InnoDB mutex will not go to sleep unless it is guaranteed to be
woken up again by another thread currently holding the mutex. This made
possible a race where a thread could get stuck waiting for a mutex that is in
fact no longer locked. If that thread was also holding other critical locks,
this could stall the entire server. There is an error monitor thread than can
break the stall, it runs about once per second. But if the error monitor
thread itself got stuck or was not running, then the entire server could hang
infinitely.
This was introduced on i386/amd64 platforms in 5.5.40 and 10.0.13 by an
incorrect patch that tried to fix the similar problem for PowerPC.
This commit reverts the incorrect PowerPC patch, and instead implements a fix
for PowerPC that does not change i386/amd64 behaviour, making PowerPC work
similarly to i386/amd64.
Analysis: InnoDB error monitor is responsible to call every second
sync_arr_wake_threads_if_sema_free() to wake up possible hanging
threads if they are missed in mutex_signal_object. This is not
possible if error monitor itself is on mutex/semaphore wait. We
should avoid all unnecessary mutex/semaphore waits on error monitor.
Currently error monitor calls function buf_flush_stat_update()
that calls log_get_lsn() function and there we will try to get
log_sys mutex. Better, solution for error monitor is that in
buf_flush_stat_update() we will try to get lsn with
mutex_enter_nowait() and if we did not get mutex do not update
the stats.
Fix: Use log_get_lsn_nowait() function on buf_flush_stat_update()
function. If returned lsn is 0, we do not update flush stats.
log_get_lsn_nowait() will use mutex_enter_nowait() and if
we get mutex we return a correct lsn if not we return 0.
The debug configuration parameter innodb_optimistic_insert_debug
which was introduced for testing corner cases in B-tree handling
had a bug in it. The value 1 would trigger an infinite sequence
of page splits.
Fix: When the value 1 is specified, disable this debug feature.
Approved by Yasufumi Kinoshita
The debug configuration parameter innodb_optimistic_insert_debug
which was introduced for testing corner cases in B-tree handling
had a bug in it. The value 1 would trigger an infinite sequence
of page splits.
Fix: When the value 1 is specified, disable this debug feature.
Approved by Yasufumi Kinoshita
FROM A FUNCTION
Scenario:
In a stored procedure, CREATE TABLE statement is not allowed. But an
exception is provided for CREATE TEMPORARY TABLE. We can create a temporary
table in a stored procedure.
Let there be two stored functions f1 and f2 and two stored procedures p1 and
p2. Their properties are as follows:
. stored function f1() calls stored procedure p1().
. stored function f2() calls stored procedure p2().
. stored procedure p1() creates temporary table t1.
. stored procedure p2() does DML on t1.
Consider the following situation:
1. Autocommit mode is on.
2. select f1()
3. select f2()
Step 2: In this step, t1 would be created via p1(). A table level transaction
lock would have been taken. The ::external_lock() would not have been called
on this table. At the end of step 2, because of autocommit mode on, this table
level lock will be released.
Step 3: When we execute DML on table t1 via p2() we have two problems:
Problem 1:
The function ha_innobase::external_lock() would have been called but since
it is a select query no table level locks would have been taken. Hence the
following assert will fail:
ut_ad(lock_table_has(thr_get_trx(thr), index->table, LOCK_IX));
Solution:
The solution would be to identify this situation and take a table level lock
and use the proper lock type prebuilt->select_lock_type = LOCK_X for DML
operations.
Problem 2:
Another problem is that in step 3, ha_innobase::open() is never called on
the table t1.
Solution:
The solution would be to identify this situation and call re-init the handler
of table t1.
rb#6429 approved by Krunal.
FROM A FUNCTION
Scenario:
In a stored procedure, CREATE TABLE statement is not allowed. But an
exception is provided for CREATE TEMPORARY TABLE. We can create a temporary
table in a stored procedure.
Let there be two stored functions f1 and f2 and two stored procedures p1 and
p2. Their properties are as follows:
. stored function f1() calls stored procedure p1().
. stored function f2() calls stored procedure p2().
. stored procedure p1() creates temporary table t1.
. stored procedure p2() does DML on t1.
Consider the following situation:
1. Autocommit mode is on.
2. select f1()
3. select f2()
Step 2: In this step, t1 would be created via p1(). A table level transaction
lock would have been taken. The ::external_lock() would not have been called
on this table. At the end of step 2, because of autocommit mode on, this table
level lock will be released.
Step 3: When we execute DML on table t1 via p2() we have two problems:
Problem 1:
The function ha_innobase::external_lock() would have been called but since
it is a select query no table level locks would have been taken. Hence the
following assert will fail:
ut_ad(lock_table_has(thr_get_trx(thr), index->table, LOCK_IX));
Solution:
The solution would be to identify this situation and take a table level lock
and use the proper lock type prebuilt->select_lock_type = LOCK_X for DML
operations.
Problem 2:
Another problem is that in step 3, ha_innobase::open() is never called on
the table t1.
Solution:
The solution would be to identify this situation and call re-init the handler
of table t1.
rb#6429 approved by Krunal.
MDEV-6483 - Deadlock around rw_lock_debug_mutex on PPC64
This problem affects only debug builds on PPC64.
There are at least two race conditions around
rw_lock_debug_mutex_enter and rw_lock_debug_mutex_exit:
- rw_lock_debug_waiters was loaded/stored without setting
appropriate locks/memory barriers.
- there is a gap between calls to os_event_reset() and
os_event_wait() and in such case we're supposed to pass
return value of the former to the latter.
Fixed by replacing self-cooked spinlocks with system mutexes.
These days system mutexes offer much better performance. OTOH
performance is not that critical for debug builds.
MDEV-6450 - MariaDB crash on Power8 when built with advance tool
chain
InnoDB mutex_exit() function calls __sync_test_and_set() to release
the lock. According to manual this function is supposed to create
"acquire" memory barrier whereas in fact we need "release" memory
barrier at mutex_exit().
The problem isn't repeatable with gcc because it creates
"acquire-release" memory barrier for __sync_test_and_set().
ATC creates just "acquire" barrier.
Fixed by creating proper barrier at mutex_exit() by using
__sync_lock_release() instead of __sync_test_and_set().
Part of this work is based on Stewart Smitch's memory barrier and lower priori
patches for power8.
- Added memory syncronization for innodb & xtradb for power8.
- Added HAVE_WINDOWS_MM_FENCE to CMakeList.txt
- Added os_isync to fix a syncronization problem on power
- Added log_get_lsn_nowait which is now used srv_error_monitor_thread to ensur
if log mutex is locked.
All changes done both for InnoDB and Xtradb
~40% bugfixed(*) applied
~40$ bugfixed reverted (incorrect or we're not buggy)
~20% bugfixed applied, despite us being not buggy
(*) only changes in the server code, e.g. not cmakefiles
SLOW/CRASHES SEMAPHORE
Problem:
There are 2 lakh tables - fk_000001, fk_000002 ... fk_200000. All of them
are related to the same parent_table through a foreign key constraint.
When the parent_table is loaded into the dictionary cache, all the child table
will also be loaded. This is taking lot of time. Since this operation happens
when the dictionary latch is taken, the scenario leads to "long semaphore wait"
situation and the server gets killed.
Analysis:
A simple performance analysis showed that the slowness is because of the
dict_foreign_find() function. It does a linear search on two linked list
table->foreign_list and table->referenced_list, looking for a particular
foreign key object based on foreign->id as the key. This is called two
times for each foreign key object.
Solution:
Introduce a rb tree in table->foreign_rbt and table->referenced_rbt, which
are some sort of index on table->foreign_list and table->referenced_list
respectively, using foreign->id as the key. These rbt structures will be
solely used by dict_foreign_find().
rb#5599 approved by Vasil
SLOW/CRASHES SEMAPHORE
Problem:
There are 2 lakh tables - fk_000001, fk_000002 ... fk_200000. All of them
are related to the same parent_table through a foreign key constraint.
When the parent_table is loaded into the dictionary cache, all the child table
will also be loaded. This is taking lot of time. Since this operation happens
when the dictionary latch is taken, the scenario leads to "long semaphore wait"
situation and the server gets killed.
Analysis:
A simple performance analysis showed that the slowness is because of the
dict_foreign_find() function. It does a linear search on two linked list
table->foreign_list and table->referenced_list, looking for a particular
foreign key object based on foreign->id as the key. This is called two
times for each foreign key object.
Solution:
Introduce a rb tree in table->foreign_rbt and table->referenced_rbt, which
are some sort of index on table->foreign_list and table->referenced_list
respectively, using foreign->id as the key. These rbt structures will be
solely used by dict_foreign_find().
rb#5599 approved by Vasil
This is not yet a fix. This is change to print additional information at the point
when this assertion is going to happen. Print as much information about the pages
and index to find out why next page is not a compact format.
Regression from bug#14621190 due to disabled optimistic restoration
of cursor, which required full key lookup instead of verifying
if previously positioned btree cursor could be reused.
Fixed by enable optimistic restore and adjust cursor afterward.
rb#3324 approved by Marko.
Regression from bug#14621190 due to disabled optimistic restoration
of cursor, which required full key lookup instead of verifying
if previously positioned btree cursor could be reused.
Fixed by enable optimistic restore and adjust cursor afterward.
rb#3324 approved by Marko.
--Implemented CHECK TABLE...QUICK.
Introduce CHECK TABLE...QUICK that would skip the btr_validate_index()
and btr_search_validate() call, and count the no. of records in each index.
Approved by Marko and Kevin. (rb#3567).
--Implemented CHECK TABLE...QUICK.
Introduce CHECK TABLE...QUICK that would skip the btr_validate_index()
and btr_search_validate() call, and count the no. of records in each index.
Approved by Marko and Kevin. (rb#3567).
The testcase for this bug fails randomly due to two reasons.
1. Due to ibuf merge happening background
2. Due to dict stats update which brings the evicted page back into
buffer pool.
Fix ibuf_contract_ext() to not do any merges with ibuf_debug enabled and
also changed dict_stats_update() to return fake statistics without
bringing the secondary index pages into buffer pool.
Approved by Marko. rb#3419
The testcase for this bug fails randomly due to two reasons.
1. Due to ibuf merge happening background
2. Due to dict stats update which brings the evicted page back into
buffer pool.
Fix ibuf_contract_ext() to not do any merges with ibuf_debug enabled and
also changed dict_stats_update() to return fake statistics without
bringing the secondary index pages into buffer pool.
Approved by Marko. rb#3419
