Problem is that FLUSH TABLES WITH READ LOCK first blocks threads from
starting new commits, then waits for running commits to complete. But
in-order parallel replication needs commits to happen in a particular
order, so this can easily deadlock.
To fix this problem, this patch introduces a way to temporarily pause
the parallel replication worker threads. Before starting FTWRL, we let
all worker threads complete in-progress transactions, and then
wait. Then we proceed to take the global read lock. Once the lock is
obtained, we unpause the worker threads. Now commits are blocked from
starting by the global read lock, so the deadlock will no longer occur.
EXPLAIN INSERT ... SELECT tried to use SELECT's execution path. This
caused a collection of problems:
- SELECT_DESCRIBE flag was not put into select_lex->options, which
means it was not in JOIN::select_options either (except for the first
member of the UNION).
- This caused UNION members to be executed. They would attempt to write
join output rows to the output.
- (Actual cause of the crash) second join sibling would call
result->send_eof() when finished execution. Then,
Explain_query::print_explain would attempt to write to query output
again, and cause an assertion due to non-empty query output.
[EXPLAIN] INSERT .. SELECT creates a select_insert object.
select_insert calls handler->start_bulk_insert() during
initialization.
For MyISAM/Aria this requires that a matching call to
handler->end_bulk_insert() call is made.
Regular INSERT .. SELECT accomplishes this by calling either
select_result->send_eof() or select_result->abort_result_set().
EXPLAIN INSERT ... SELECT didn't call either, which resulted in
improper de-initializaiton of handler object. Make it call
abort_result_set(), which invokes handler->end_bulk_insert().
Before this, FLUSH TABLES and FLUSH TABLES t1 (REFRESH_TABLES hereafter) were executed locally,
which also had implications for GTID. MySQL async replication does
replicate FLUSH TABLES.
Hence, this commit
a) Adds REFRESH_TABLES to TOI.
b) Skips LOCK-based REFRESH_TABLES from TOI, ie. FLUSH TABLES WITH READ
LOCK and FLUSH TABLES FOR EXPORT.
c) Executes them after reload_acl_and_cache. This is done since
reload_acl_and_cache checks for thd->locked_tables_mode and does upgrade
of MDL locks accordingly. A failure here means REFRESH_TABLES cannot
run.
i) This is also important since LOCK TABLES is not replicated in
galera.
d) Test galera_flush has been modified for
i) wsrep_replicate_myisam SESSION variable.
ii) Testing REFRESH_TABLES.
iii) Testing REFRESH_TABLES + LOCK TABLES.
e) Use different WSREP_TO_ISOLATION_BEGIN for global FLUSH TABLES and FLUSH TABLES <table,>.
(MDEV-8617: Post-fix for 10.1)
* Reset THD's PS members before returning when node is
not ready
* Add CF_SKIP_WSREP_CHECK flag to COM_STMT_XXX commands
* Skip TO replication of COM_STMT_PREPAREs for MyISAM
* Updated tests
- Part 3: Adding mem_root to push_back() and push_front()
Other things:
- Added THD as an argument to some partition functions.
- Added memory overflow checking for XML tag's in read_xml()
- Added mem_root to all calls to new Item
- Added private method operator new(size_t size) to Item to ensure that
we always use a mem_root when creating an item.
This saves use once call to current_thd per Item creation
Added mandatory thd parameter to Item (and all derivative classes) constructor.
Added thd parameter to all routines that may create items.
Also removed "current_thd" from Item::Item. This reduced number of
pthread_getspecific() calls from 290 to 177 per OLTP RO transaction.
In galera cluster, when myisam replication is enabled
(wsrep_replicate_myisam=ON), DML statements are replicated
in open_tables(). However, in case of prepared statements,
for an INSERT, open_tables() gets invoked twice. Once for
COM_STMT_PREPARE (to validate and prepare INSERT) and later
for COM_STMT_EXECUTE. As a result, the command gets replicated
twice. Same happens for REPLACE, UPDATE and DELETE commands.
Fixed by adding a check to not replicate during 'prepare'
phase. Also changed the order of conditions to make it more
efficient. Lastly, in order to support wsrep_dirty_reads, made
changes to allow COM_STMT_XXX commands to continue past initial
check even when wsrep is not ready.
RENAME TABLE, unlike other DDLs, was getting replicated before
the access check was performed. As a result, the command could
get get replicated and thus executed on other nodes, even if it
fails on the originating node due to permission issues. Fixed by
moving the logic to check user privileges before replicating the
command.
THD::>save_prep_leaf_list was set to true by multi-table update
statements with mergeable selects and never reset.
Make every statement reset it at start.
THD::>save_prep_leaf_list was set to true by multi-table update
statements with mergeable selects and never reset.
Make every statement reset it at start.
[Attempt #] Make the code that handles "Prepare" phase for multi-table
UPDATE statements handle non-merged semijoins. It can encounter them when
a prepared statement is executed for the second time.
DATABASE WHEN USING TABLE ALIASES
Issue:
-----
When using table aliases for deleting, MySQL checks
privileges against the current database and not the
privileges on the actual table or database the table
resides.
SOLUTION:
---------
While checking privileges for multi-deletes,
correspondent_table should be used since it points to the
correct table and database.
Problem :
---------
The specific issue reported in this bug is with range/list column
value that is allocated and initialized by evaluating partition
expression(item tree) during execution. After evaluation the range
list value is marked fixed [part_column_list_val]. During next
execution, we don't re-evaluate the expression and use the old value
since it is marked fixed.
Solution :
----------
One way to solve the issue is to mark all column values as not fixed
during clone so that the expression is always re-evaluated once we
attempt partition_info::fix_column_value_functions() after cloning
the part_info object during execution of DDL on partitioned table.
Reviewed-by: Jimmy Yang <Jimmy.Yang@oracle.com>
Reviewed-by: Mattias Jonsson <mattias.jonsson@oracle.com>
RB: 9424
- Changed ER(ER_...) to ER_THD(thd, ER_...) when thd was known or if there was many calls to current_thd in the same function.
- Changed ER(ER_..) to ER_THD_OR_DEFAULT(current_thd, ER...) in some places where current_thd is not necessary defined.
- Removing calls to current_thd when we have access to thd
Part of this is optimization (not calling current_thd when not needed),
but part is bug fixing for error condition when current_thd is not defined
(For example on startup and end of mysqld)
Notable renames done as otherwise a lot of functions would have to be changed:
- In JOIN structure renamed:
examined_rows -> join_examined_rows
record_count -> join_record_count
- In Field, renamed new_field() to make_new_field()
Other things:
- Added DBUG_ASSERT(thd == tmp_thd) in Item_singlerow_subselect() just to be safe.
- Removed old 'tab' prefix in JOIN_TAB::save_explain_data() and use members directly
- Added 'thd' as argument to a few functions to avoid calling current_thd.
Pretend that CREATE TABLE and CREATE TEMPORARY TABLE are
two different commands internally. The user doesn't need
to know that they both are SQLCOM_CREATE_TABLE.
Same for DROP [TEMPORARY] TABLE
Initialize abs_timeout when it is about to be used. This saves one my_hrtime()
call on hot path (when we acquire MDL lock without waiting).
When filling I_S.PROCESSLIST use THD::start_utime/THD::utime_after_query instead
of THD::start_time. This allows us to save 2 clock_gettime() calls.
Overhead change:
__clock_gettime 0.13% -> 0.11% (122 -> 76 calls per OLTP RO transaction)
my_interval_timer 0.07% -> 0.06%
my_hrtime 0.04% -> 0.01%
