When QUICK_GROUP_MIN_MAX_SELECT is initialized or being reset
it stores the prefix of the last group of the index chosen for
retrieving data (last_value). Later, when looping through records
at get_next() method, the server checks whether the retrieved
group is the last, and if so, it finishes processing.
At the same time, it looks like there is no need for that additional
check since method next_prefix() returns HA_ERR_KEY_NOT_FOUND
or HA_ERR_END_OF_FILE when there are no more satisfying records.
If we do not perform the check, we do not need to retrieve and
store last_value either.
This commit removes using of last_value from QUICK_GROUP_MIN_MAX_SELECT.
Reviewer: Sergei Petrunia <sergey@mariadb.com>
Problem:
========
- InnoDB should have two keys on the same column for the self
referencing foreign key relation.
Solution:
=========
- Allow self referential foreign key relation to work with one
key.
The binlog.flashback test tries to run using the stmt and mix
binlog_format combinations (because it includes have_log_bin.inc),
yet are meaningless tests because the server starts with the
--flashback option, forcing binlog_format=row.
This patch optimizes test time by removing the stmt and mix format
combinations from running.
Reviewed By:
============
Andrei Elkin <andrei.elkin@mariadb.com>
InnoDB could return off-by-1 estimates for the involved tables.
This would cause off-by-many difference in join output cardinality
for the top-level SELECT, and so different query plan for the subquery.
The fix: Introduce mysql-test/include/innodb_stable_estimates.{inc,opt}
which disables InnoDB's background statistics collection, and use it.
This test happens to fail if it runs after test
galera_inject_bf_long_wait.
And the reason is that galera_bf_lock_wait greps for message
"BF lock wait long" in the error log, and expects that grep matches
no lines. Whereas galera_inject_bf_long_wait intentionally causes the
message to appear in the log. The fix consists in using
assert_grep.inc with option assert_only_after, such that
galera_bf_lock_wait is limited to grep only those lines that appeared
in the log after it started to execute.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Extending the rpl_mysqldump_slave.test to incorporate the
delete-master-logs option, and this options is alias of
get binlogs: show master status -> flush logs -> purge binary logs to <new_binlog>
sequence and this test is derived using the same pattern.
also we measure the pre and post log state
on the master following the mysqldump process. Introducing
assertions to validate the correct state.
Using mysql.slow_log was a test table would generate more than
one row if there was more than one row in the table.
Replace this table with a empty table with PK.
Reviewer: Rex Johnston
row_upd_clust_rec_by_insert(): If we are resuming from a lock wait,
reset the 'disowned' flag of the BLOB pointers in 'entry' that we
copied from 'rec' on which we had invoked btr_cur_disown_inherited_fields()
before the lock wait started. In this way, the inserted record with
the updated PRIMARY KEY value will have the BLOB ownership associated
with itself, like it is supposed to be.
Note: If the lock wait had been aborted, then rollback would have
invoked btr_cur_unmark_extern_fields() and no corruption would be possible.
Reviewed by: Vladislav Lesin
Tested by: Matthias Leich
- Added the parameter stats_persistent=0 for InnoDB engine.
- Before printing metadata_lock_info query, make sure that
InnoDB does complete purging.
Reviewed by: Marko Mäkelä
trx_t::commit_in_memory(): Empty the detailed_error string, so that
FOREIGN KEY error messages from an earlier transaction will not be
wrongly reused in ha_innobase::get_error_message().
Reviewed by: Thirunarayanan Balathandayuthapani
It may be the case that for some reason, -Werror deprecated
for instance, that mallinfo isn't detected. In this case the
malloc.h headers won't be included which defined the mallinfo2
function and its structure.
Re-organise so that either function pulls in the header.
The problem is that background statistics can race with statistics update
during INSERT and cause slightly inaccurate `Rows` count in table statistics
(this is deliberate to avoid excessive locking overhead). This was seen as
occasional .result difference in the test.
Mask out the unstable `Rows` column from SHOW TABLE STATUS; the value is not
related to what is being tested in this part of the test case.
Run ANALYZE TABLE before SHOW EXPLAIN to get stable row count in output.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
Fix some random test failures following MDEV-32168 push.
Don't blindly set $rpl_only_running_threads in many places. Instead explicit
stop only the IO or SQL thread, as appropriate. Setting it interfered with
rpl_end.inc in some cases. Rather than clearing it afterwards, better to
not set it at all when it is not needed, removing ambiguity in the test
about the state of the replication threads.
Don't fail the test if include/stop_slave_io.inc finds an error in the IO
thread after stop. Such errors can be simply because slave stop happened in
the middle of the IO thread's initial communication with the master.
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
- Commit fc31e3114b2538d152194d17ff0f0439db565634(MDEV-8179) doesn't
report the progress of inplace alter completely. It just does only
in row_merge_sort(). Removing the progress report function completely
This avoids the scenario in MDEV-32849, when the unlock happens after
the connection has been freed, say in rollback. This is done in 10.5+
after the commit a26700cca5.
It may or may not prevent potential other scenarios where spider has
locked something, then for some reason the statement needs to be
rolled back and spider frees the connection, and then spider proceeds
to use the freed connection. But at least we fix the regression
introduced by MDEV-30014 to 10.4 and bring 10.4 closer in parity with
10.5+.
Other things:
- Added DBUG_EXECUTE_IF("print_allocated_thread_memory") at end of query
to easier find not freed memory allocated by THD
- Removed free_root() from plugin_init() that did nothing.
The leaks are all 40 bytes and happens in this call stack when running
mtr vcol.vcol_syntax:
alloc_root()
...
Virtual_column_info::fix_and_check_exp()
...
Delayed_insert::get_local_table()
The problem was that one copied a MEM_ROOT from THD to a TABLE without
taking into account that new blocks would be allocated through the
TABLE memroot (and would thus be leaked).
In general, one should NEVER copy MEM_ROOT from one object to another
without clearing the copied memroot!
Fixed by, at end of get_local_table(), copy all new allocated objects
to client_thd->mem_root.
Other things:
- Removed references to MEM_ROOT::total_alloc that was wrongly left
after a previous commit
The reason for this change are the following:
- If we call set_killed() from one thread to kill another thread with
a message, there may be concurrent usage of the MEM_ROOT which is
not supported (this could cause memory corruption).
We do not currently have code that does this, but the API allows this
and it is better to be fix the issue before it happens.
- The per thread memory tracking does not work if one thread uses
another threads MEM_ROOT.
- set_killed() can be called if a MEM_ROOT allocation fails. In this case
it is not good to try to allocate more memory from potentially the same
MEM_ROOT.
Fix is to use my_malloc() instead of mem_root for killed messages.
The stack function trace for this bug is:
libc
my_free
free_root
acl_reload
The crash happens because acl_memroot gets corrupted.
The issue was that during FLUSH PRIVILEGES we discard the old
privileges and create new ones. We have protection in place that no
one can accesses the privileges during this time.
However one short piece of code called during login of a new user, or
change password, was not properly protected, which could in some very
rare circumstances case a memory overwrite of a MEMROOT object if
at the same time another thread calls FLUSH PRIVILEGES.
This it issue is fixed by adding protection around set_user_salt().
I also added asserts to other code that is using the acl_memroot to
ensure that it is properly proteced everywhere.
- Record unrecorded tests from `rpl` suite to `engines/funcs` suite
(done by d8e448ba1b):
1) Record test `rpl_row_until` from commit d95fa7e332
2) Record test `rpl_slave_status` from commit a7d186a17d
- Stop only running threads for `engines/funcs.rpl_server_id1.test` that
is not the same as `rpl.rpl_server_id1.test`
- Reviewer: <knielsen@knielsen-hq.org>
<andrei.elkin@mariadb.com>
mysql_prepare_alter_table(): Alter table should check whether
foreign key exists when it expected to exists and
report the error in early stage
dict_foreign_parse_drop_constraints(): Don't throw error if the
foreign key constraints doesn't exist when if exists is given
in the statement.
move MYSQL::fields down, replacing MYSQL::unused5
this way only MYSQL::fields and MYSQL::field_alloc will still have
different offset in C/C and the server, but all other MYSQL members
will get back in sync.
luckily, plugins shouldn't need MYSQL::fields or MYSQL::field_alloc
added a check to ensure both MYSQL structures are always of
the same size.
After successful connection, server always sets SERVER_STATUS_AUTOCOMMIT
in server_status in the OK packet. This is wrong, if global variable
autocommit=0.
Fixed THD::init(), added mysql_client_test test.
Thanks to Diego Dupin for the providing the patch.
Signed-off-by: Vladislav Vaintroub <vvaintroub@gmail.com>
This bug was caused by a patch for the task MDEV-32733.
Incorrect memory root was used for allocation of memory
pointed by the data memebr Item_func_json_contains_path::p_found.
Let us make the test compatible with ./mtr --repeat
and convert variable_value to integer, so that comparisons like
16>9 will work as intended, instead of being compared as '16'<'9'.
The bug is fixed by the patch ported from MySQL. See the comprehensive
description below.
commit 455c4e8810c76430719b1a08a63ca0f69f44678a
Author: Guilhem Bichot <guilhem.bichot@oracle.com>
Date: Fri Mar 13 17:51:27 2015 +0100
Bug#17668844: CRASH/ASSERT AT ITEM_TYPE_HOLDER::VAL_STR IN ITEM.C
We have a predicate of the form:
literal_row <=> (a UNION)
The subquery is constant, so Item_cache objects are used for its
SELECT list.
In order, this happens:
- Item_subselect::fix_fields() calls select_lex_unit::prepare,
where we create Item_type_holder's
(appended to unit->types list), create the tmp table (using type info
found in unit->types), and call fill_item_list() to put the
Item_field's of this table into unit->item_list.
- Item_subselect::fix_length_and_dec() calls set_row() which
makes Item_cache's of the subquery wrap the Item_type_holder's
- When/if a first result row is found for the subquery,
Item_cache's are re-pointed to unit->item_list
(i.e. Item_field objects which reference the UNION's tmp table
columns) (see call to Item_singlerow_subselect::store()).
- In our subquery, no result row is found, so the Item_cache's
still wrap Item_type_holder's; evaluating '<=>' reads the
value of those, but Item_type_holder objects are not expected to be
evaluated.
Fix: instead of putting unit->types into Item_cache, and later
replacing with unit->item_list, put unit->item_list in Item_cache from
the start.
Approved by Oleksandr Byelkin <sanja@mariadb.com>
This is the follow-up patch that removes explicit use of thd->stmt_arena
for memory allocation and replaces it with call of the method
THD::active_stmt_arena_to_use()
Additionally, this patch adds extra DBUG_ASSERT to check that right
query arena is in use.
This patch is actually follow-up for the task
MDEV-23902: MariaDB crash on calling function
to use correct query arena for a statement. In case invocation of
a function is in progress use its call arena, else use current
query arena that can be either a statement or a regular query arena.
We remove the call to update spider persistent table stats (sts/crd)
in spider_free_share(). This prevents spider from opening and closing
further tables during close(), which fixes the following issues:
MDEV-28739: ha_spider::close() is called during tdc_start_shutdown(),
which is called after query_cache_destroy(). Closing the sts/crd Aria
tables will trigger a call to Query_cache::invalidate_table(), which
will attempt to use the query cache mutex structure_guard_mutex
destroyed previously.
MDEV-29421: during ha_spider::close(), spider_free_share() could
trigger another spider_free_share() through updating sts/crd table,
because open_table() calls tc_add_table(), which could trigger another
ha_spider::close()...
Since spider sts/crd system tables are only updated here, there's no
use for these tables any more, and we remove all uses of these tables
too.
The removal should not cause any performance issue, as in memory
spider table stats are only updated based on a time
interval (spider_sts_interval and spider_crd_interval), which defaults
to 10 seconds. It should not affect accuracy either, due to the
infrequency of server restart. And inaccurate stats are not a problem
for optimizer anyway.
To be on the safe side, we defer the removal of the spider sts/crd
tables themselves to future.
btr_cur_update_in_place(): Update the DB_TRX_ID,DB_ROLL_PTR also on the
compressed copy of the page. In a test case, a server built with
cmake -DWITH_INNODB_EXTRA_DEBUG=ON would crash in page_zip_validate()
due to the inconsistency. In a normal debug build, a different assertion
would fail, depending on when the uncompressed page was restored from
the compressed page.
In MariaDB Server 10.5, this bug had already been fixed by
commit b3d02a1fcf (MDEV-12353).
