The previous commit for fixing MDEV-35446 disabled setting
Galera errors on COM_STMT_PREPARE commands.
As a side effect, a number of tests were started to fail
due to the client receiving different error codes from the
ones expected in the test dependending on whether --ps-protocol
was used.
Also, in the case of test galera_ftwrl, it was found that
it is expected that during COM_STMT_PREPARE command, we
may perform a sync wait operation, which can fail with
LOCK_WAIT_TIMEOUT error.
The revised fix consists in anticipating the call to
wsrep_after_command_before_result(), so that we check for
BF aborts or errors during statement prepare, before sending
back the statement metadata message to client.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Test failed sporadically when --ps-protocol was enabled:
a transaction that was BF aborted on COMMIT would succeed
instead of reporting the expected deadlock error.
The reason for the failure was that, depending on timing,
the transaction was BF aborted while the COMMIT statement
was being prepared through a COM_STMT_PREPARE command.
In the failing cases, the transaction was BF aborted
after COM_STMT_PREPARE had already disabled the diagnostics
area of the client. Attempt to override the deadlock error
towards the end of dispatch_command() would be skipped,
resulting in a successful COMMIT even if the transaction
is aborted.
This bug affected the following MTR tests:
- galera_insert_multi
- galera_nopk_unicode
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
The problem was that when using clang + asan, we do not get a correct value
for the thread stack as some local variables are not allocated at the
normal stack.
It looks like that for example clang 18.1.3, when compiling with
-O2 -fsanitize=addressan it puts local variables and things allocated by
alloca() in other areas than on the stack.
The following code shows the issue
Thread 6 "mariadbd" hit Breakpoint 3, do_handle_one_connection
(connect=0x5080000027b8,
put_in_cache=<optimized out>) at sql/sql_connect.cc:1399
THD *thd;
1399 thd->thread_stack= (char*) &thd;
(gdb) p &thd
(THD **) 0x7fffedee7060
(gdb) p $sp
(void *) 0x7fffef4e7bc0
The address of thd is 24M away from the stack pointer
(gdb) info reg
...
rsp 0x7fffef4e7bc0 0x7fffef4e7bc0
...
r13 0x7fffedee7060 140737185214560
r13 is pointing to the address of the thd. Probably some kind of
"local stack" used by the sanitizer
I have verified this with gdb on a recursive call that calls alloca()
in a loop. In this case all objects was stored in a local heap,
not on the stack.
To solve this issue in a portable way, I have added two functions:
my_get_stack_pointer() returns the address of the current stack pointer.
The code is using asm instructions for intel 32/64 bit, powerpc,
arm 32/64 bit and sparc 32/64 bit.
Supported compilers are gcc, clang and MSVC.
For MSVC 64 bit we are using _AddressOfReturnAddress()
As a fallback for other compilers/arch we use the address of a local
variable.
my_get_stack_bounds() that will return the address of the base stack
and stack size using pthread_attr_getstack() or NtCurrentTed() with
fallback to using the address of a local variable and user provided
stack size.
Server changes are:
- Moving setting of thread_stack to THD::store_globals() using
my_get_stack_bounds().
- Removing setting of thd->thread_stack, except in functions that
allocates a lot on the stack before calling store_globals(). When
using estimates for stack start, we reduce stack_size with
MY_STACK_SAFE_MARGIN (8192) to take into account the stack used
before calling store_globals().
I also added a unittest, stack_allocation-t, to verify the new code.
Reviewed-by: Sergei Golubchik <serg@mariadb.org>
Added new test scenario in galera.galera_bf_kill
test to make the issue surface. The tetst scenario has
a multi statement transaction containing a KILL command.
When the KILL is submitted, another transaction is
replicated, which causes BF abort for the KILL command
processing. Handling BF abort rollback while executing
KILL command causes node hanging, in this scenario.
sql_kill() and sql_kill_user() functions have now fix,
to perform implicit commit before starting the KILL command
execution. BEcause of the implicit commit, the KILL execution
will not happen inside transaction context anymore.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Problem was that we did not found that table was partitioned
and then we should find what is actual underlaying storage
engine.
We should not use RSU for !InnoDB tables.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Replication of MyISAM and Aria DML is experimental and best
effort only. Earlier change make INSERT SELECT on both
MyISAM and Aria to replicate using TOI and STATEMENT
replication. Replication should happen only if user
has set needed wsrep_mode setting.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Address Sanitizer's know how to detect stack overrun, so there's
no point in us doing it.
As evidenced by perfschema tests where signficant test failures
because this function failed under ASAN (MDEV-33210).
Also, so since clang-16, we cannot assume much about how local
variables are allocated on the stack (MDEV-31605).
Disabling check idea thanks to Sanja.
Rectify cases of mismatched brackets and address
possible cases of division by zero by checking if
the denominator is zero before dividing.
No functional changes were made.
All new code of the whole pull request, including one or several
files that are either new files or modified ones, are contributed
under the BSD-new license. I am contributing on behalf of my
employer Amazon Web Services, Inc.
Having Item_func_not items in item trees breaks assumptions during the
optimization phase about transformation possibilities in fix_fields().
Remove Item_func_not by extending normalization during parsing.
Reviewed by Oleksandr Byelkin (sanja@mariadb.com)
New error codes can only be added in the latest major version.
Adding ER_KILL_DENIED_HIGH_PRIORITY would shift by one all
error codes that were added in MariaDB Server 10.6 or later.
This amends commit 1001dae186
Suggested by: Sergei Golubchik
Changed error code for Galera unkillable threads to
be ER_KILL_DENIED_HIGH_PRIORITY giving message
This is a high priority thread/query and cannot be killed
without the compromising consistency of the cluster
also a warning is produced
Thread %lld is [wsrep applier|high priority] and cannot be killed
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Adding the test for the length of lex->name into show_create_db().
Without this test writes beyond the end of db_name_buff were possible
upon a too long database name.
This is regression from commit 3228c08fa8. Problem is that
when table storage engine is determined there should be
check is table partitioned and if it is then determine
partition implementing storage engine.
Reported bug is reproducible only with --log-bin so make
sure tests changed by 3228c08fa8 and new test are run
with --log-bin and binlog disabled.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Correct the second parameter for strxnmov to prevent potential buffer
overflows. The second parameter must be one less than the size of the
input buffer to avoid writing past the end of the buffer.
While the second parameter is usually correct, there are exceptions
that need fixing.
This commit addresses the issue within frm_file_exists() and other
affected places.
The patch for MDEV-31340 fixed the following bugs:
MDEV-33084 LASTVAL(t1) and LASTVAL(T1) do not work well with lower-case-table-names=0
MDEV-33085 Tables T1 and t1 do not work well with ENGINE=CSV and lower-case-table-names=0
MDEV-33086 SHOW OPEN TABLES IN DB1 -- is case insensitive with lower-case-table-names=0
MDEV-33088 Cannot create triggers in the database `MYSQL`
MDEV-33103 LOCK TABLE t1 AS t2 -- alias is not case sensitive with lower-case-table-names=0
MDEV-33108 TABLE_STATISTICS and INDEX_STATISTICS are case insensitive with lower-case-table-names=0
MDEV-33109 DROP DATABASE MYSQL -- does not drop SP with lower-case-table-names=0
MDEV-33110 HANDLER commands are case insensitive with lower-case-table-names=0
MDEV-33119 User is case insensitive in INFORMATION_SCHEMA.VIEWS
MDEV-33120 System log table names are case insensitive with lower-cast-table-names=0
Backporting the fixes from 11.5 to 10.5
- Fix to avoid mysqltest client getting killed abruptly during
mysql_shutdown(). When Galera replication is shutdown, wait for
THDs with `thd->stmt_da()->is_eof()` to disconnect (these are about
to disconnect anyway).
- Extract duplicate code from `wsrep_stop_replication()` and
`wsrep_shutdown_replication()` in a new function.
- No need to use a custom `shutdown_mysqld.inc` in galera
suite. Delete it, so that the one in `mysql-test/include/` is used.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
There were several races in the main.kill_processlist-6619 testcase:
- Lingering connections from a previous test case could be visible in SHOW
PROCESSLIST and cause .result diff.
- A sync point "dispatch_command_end" was ineffective, as it was consumed at
the end of the SET DEBUG command itself.
- The signal from sync point "before_execute_sql_command" could override an
earlier signal, causing DEBUG_SYNC timeout and test failure.
- The final SHOW PROCESSLIST could occasionally see a connection in state
"Busy" instead of the expected "Sleep".
Signed-off-by: Kristian Nielsen <knielsen@knielsen-hq.org>
In case there is a view that queried from a stored routine or
a prepared statement and this temporary table is dropped between
executions of SP/PS, then it leads to hitting an assertion
at the SELECT_LEX::fix_prepare_information. The fired assertion
was added by the commit 85f2e4f8e8
(MDEV-32466: Potential memory leak on executing of create view statement).
Firing of this assertion means memory leaking on execution of SP/PS.
Moreover, if the added assert be commented out, different result sets
can be produced by the statement SELECT * FROM the hidden table.
Both hitting the assertion and different result sets have the same root
cause. This cause is usage of temporary table's metadata after the table
itself has been dropped. To fix the issue, reload the cache of stored
routines. To do it cache of stored routines is reset at the end of
execution of the function dispatch_command(). Next time any stored routine
be called it will be loaded from the table mysql.proc. This happens inside
the method Sp_handler::sp_cache_routine where loading of a stored routine
is performed in case it missed in cache. Loading is performed unconditionally
while previously it was controlled by the parameter lookup_only. By that
reason the signature of the method Sroutine_hash_entry::sp_cache_routine
was changed by removing unused parameter lookup_only.
Clearing of sp caches affects the test main.lock_sync since it forces
opening and locking the table mysql.proc but the test assumes that each
statement locks its tables once during its execution. To keep this invariant
the debug sync points with names "before_lock_tables_takes_lock" and
"after_lock_tables_takes_lock" are not activated on handling the table
mysql.proc
Problem was that REPLACE was using consistency check that started
TOI and we tried to rollback it.
Do not use wsrep_before_rollback and wsrep_after_rollback if
we are runing consistency check because no writeset keys are
in that case added. Do not allow consistency check usage
if table storage for target table is not InnoDB, instead
give warning. REPLACE|SELECT INTO ... SELECT will use
now TOI if table storage for target table is not InnoDB
to maintain consistency between galera nodes.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
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.
- Allow database level access via `LOCK TABLES` to execute statement
`BACKUP [un]LOCK <object>`
- `BACKUP UNLOCK` works only with `RELOAD` privilege.
In case there is `LOCK TABLES` privilege without `RELOAD` privilege,
we check if backup lock is taken before.
If it is not we raise an error of missing `RELOAD` privilege.
- We had to remove any error/warnings from calling functions because
`thd->get_stmt_da()->m_status` will be set to error and will break
`my_ok()`.
- Added missing test coverage of `RELOAD` privilege to `main.grant.test`
Reviewer: <daniel@mariadb.org>
In some cases "SHOW PROCESSLIST" could show "Reset for next command"
as State, even if the previous query had finished properly.
Fixed by clearing State after end of command and also setting the State
for the "Connect" command.
Other things:
- Changed usage of 'thd->set_command(COM_SLEEP)' to
'thd->mark_connection_idle()'.
- Changed thread_state_info() to return "" instead of NULL. This is
just a safety measurement and in line with the logic of the
rest of the function.
standard table KEY_COLUMN_USAGE should only show keys where
a user has some privileges on every column of the key
standard table TABLE_CONSTRAINTS should show tables where
a user has any non-SELECT privilege on the table or on any column
of the table
standard table REFERENTIAL_CONSTRAINTS is defined in terms of
TABLE_CONSTRAINTS, so the same rule applies. If the user
has no rights to see the REFERENCED_TABLE_NAME value, it should be NULL
SHOW INDEX (and STATISTICS table) is non-standard, but it seems
reasonable to use the same logic as for KEY_COLUMN_USAGE.
The pointer was used deep in the call path.
Resolve this by setting the pointer to NULL at the end of
the function.
Tested with gcc-13.3.1 (fc38)
The warning disable 38fe266ea9 can be reverted in 10.6+ on merge.
The parser works as follows:
The rule expr_lex returns a pointer to a newly created sp_expr_lex
instance which is not linked to any MariaDB structures yet - it is
pointed only from a Bison stack variable. The sp_expr_lex instance
gets linked to other structures (such as sp_instr_jump_if_not) later,
after scanning some following grammar.
Problem before the fix:
If a parse error happened immediately after expr_lex (before it got linked),
the created sp_expr_lex value got lost causing a memory leak.
Fix:
- Using Bison's "destructor" directive to free the results of expr_lex
on parse/oom errors.
- Moving the call for LEX::cleanup_lex_after_parse_error() from
MYSQL_YYABORT and yyerror inside parse_sql().
This is needed because Bison calls destructors after yyerror(),
while it's important to delete the sp_expr_lex instance before
LEX::cleanup_lex_after_parse_error().
The latter frees the memory root containing the sp_expr_lex instance.
After this change the code block are executed in the following order:
- yyerror() -- now only raises the error to DA (no cleanup done any more)
- %destructor { delete $$; } <expr_lex> -- destructs the sp_expr_lex instance
- LEX::cleanup_lex_after_parse_error() -- frees the memory root containing
the sp_expr_lex instance
- Removing the "delete sublex" related code from restore_lex():
- restore_lex() is called in most cases on success, when delete is not needed.
- There is one place when restore_lex() is called on error:
In sp_create_assignment_instr(). But in this case LEX::sp_lex_in_use
is true anyway.
The patch adds a new DBUG_ASSERT(lex->sp_lex_in_use) to guard this.
The problem seems to be a deadlock between KILL command execution
and BF abort issued by an applier, where:
* KILL has locked victim's LOCK_thd_kill and LOCK_thd_data.
* Applier has innodb side global lock mutex and victim trx mutex.
* KILL is calling innobase_kill_query, and is blocked by innodb
global lock mutex.
* Applier is in wsrep_innobase_kill_one_trx and is blocked by
victim's LOCK_thd_kill.
The fix in this commit removes the TOI replication of KILL command
and makes KILL execution less intrusive operation. Aborting the
victim happens now by using awake_no_mutex() and ha_abort_transaction().
If the KILL happens when the transaction is committing, the
KILL operation is postponed to happen after the statement
has completed in order to avoid KILL to interrupt commit
processing.
Notable changes in this commit:
* wsrep client connections's error state may remain sticky after
client connection is closed. This error message will then pop
up for the next client session issuing first SQL statement.
This problem raised with test galera.galera_bf_kill.
The fix is to reset wsrep client error state, before a THD is
reused for next connetion.
* Release THD locks in wsrep_abort_transaction when locking
innodb mutexes. This guarantees same locking order as with applier
BF aborting.
* BF abort from MDL was changed to do BF abort on server/wsrep-lib
side first, and only then do the BF abort on InnoDB side. This
removes the need to call back from InnoDB for BF aborts which originate
from MDL and simplifies the locking.
* Removed wsrep_thd_set_wsrep_aborter() from service_wsrep.h.
The manipulation of the wsrep_aborter can be done solely on
server side. Moreover, it is now debug only variable and
could be excluded from optimized builds.
* Remove LOCK_thd_kill from wsrep_thd_LOCK/UNLOCK to allow more
fine grained locking for SR BF abort which may require locking
of victim LOCK_thd_kill. Added explicit call for
wsrep_thd_kill_LOCK/UNLOCK where appropriate.
* Wsrep-lib was updated to version which allows external
locking for BF abort calls.
Changes to MTR tests:
* Disable galera_bf_abort_group_commit. This test is going to
be removed (MDEV-30855).
* Record galera_gcache_recover_manytrx as result file was incomplete.
Trivial change.
* Make galera_create_table_as_select more deterministic:
Wait until CTAS execution has reached MDL wait for multi-master
conflict case. Expected error from multi-master conflict is
ER_QUERY_INTERRUPTED. This is because CTAS does not yet have open
wsrep transaction when it is waiting for MDL, query gets interrupted
instead of BF aborted. This should be addressed in separate task.
* A new test galera_kill_group_commit to verify correct behavior
when KILL is executed while the transaction is committing.
Co-authored-by: Seppo Jaakola <seppo.jaakola@iki.fi>
Co-authored-by: Jan Lindström <jan.lindstrom@galeracluster.com>
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
This is a backport from 10.5.
The problem seems to be a deadlock between KILL command execution
and BF abort issued by an applier, where:
* KILL has locked victim's LOCK_thd_kill and LOCK_thd_data.
* Applier has innodb side global lock mutex and victim trx mutex.
* KILL is calling innobase_kill_query, and is blocked by innodb
global lock mutex.
* Applier is in wsrep_innobase_kill_one_trx and is blocked by
victim's LOCK_thd_kill.
The fix in this commit removes the TOI replication of KILL command
and makes KILL execution less intrusive operation. Aborting the
victim happens now by using awake_no_mutex() and ha_abort_transaction().
If the KILL happens when the transaction is committing, the
KILL operation is postponed to happen after the statement
has completed in order to avoid KILL to interrupt commit
processing.
Notable changes in this commit:
* wsrep client connections's error state may remain sticky after
client connection is closed. This error message will then pop
up for the next client session issuing first SQL statement.
This problem raised with test galera.galera_bf_kill.
The fix is to reset wsrep client error state, before a THD is
reused for next connetion.
* Release THD locks in wsrep_abort_transaction when locking
innodb mutexes. This guarantees same locking order as with applier
BF aborting.
* BF abort from MDL was changed to do BF abort on server/wsrep-lib
side first, and only then do the BF abort on InnoDB side. This
removes the need to call back from InnoDB for BF aborts which originate
from MDL and simplifies the locking.
* Removed wsrep_thd_set_wsrep_aborter() from service_wsrep.h.
The manipulation of the wsrep_aborter can be done solely on
server side. Moreover, it is now debug only variable and
could be excluded from optimized builds.
* Remove LOCK_thd_kill from wsrep_thd_LOCK/UNLOCK to allow more
fine grained locking for SR BF abort which may require locking
of victim LOCK_thd_kill. Added explicit call for
wsrep_thd_kill_LOCK/UNLOCK where appropriate.
* Wsrep-lib was updated to version which allows external
locking for BF abort calls.
Changes to MTR tests:
* Disable galera_bf_abort_group_commit. This test is going to
be removed (MDEV-30855).
* Record galera_gcache_recover_manytrx as result file was incomplete.
Trivial change.
* Make galera_create_table_as_select more deterministic:
Wait until CTAS execution has reached MDL wait for multi-master
conflict case. Expected error from multi-master conflict is
ER_QUERY_INTERRUPTED. This is because CTAS does not yet have open
wsrep transaction when it is waiting for MDL, query gets interrupted
instead of BF aborted. This should be addressed in separate task.
* A new test galera_kill_group_commit to verify correct behavior
when KILL is executed while the transaction is committing.
Co-authored-by: Seppo Jaakola <seppo.jaakola@iki.fi>
Co-authored-by: Jan Lindström <jan.lindstrom@galeracluster.com>
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Problem for Galera is the fact that sequences are not really
transactional. Sequence operation is committed immediately
in sql_sequence.cd and later Galera could find out that
we have changes but actual statement is not there anymore.
Therefore, we must make some restrictions what kind
of sequences Galera can support.
(1) Galera cluster supports only sequences implemented
by InnoDB storage engine. This is because Galera replication
supports currently only InnoDB.
(2) We do not allow LOCK TABLE on sequence object and
we do not allow sequence creation under LOCK TABLE, instead
lock is released and we issue warning.
(3) We allow sequences with NOCACHE definition or with
INCREMEMENT BY 0 CACHE=n definition. This makes sure that
sequence values are unique accross Galera cluster.
Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
Assertion `thd->mdl_context.is_lock_owner()` fires when a client is
disconnected, while transaction and and a table is opened through
`HANDLER` interface.
Reason for the assertion is that when a connection closes, its ongoing
transaction is eventually rolled back in
`Wsrep_client_state::bf_rollback()`. This method also releases explicit
which are expected to survive beyond the transaction lifetime.
This patch also removes calls to `mysql_ull_cleanup()`. User level
locks are not supported in combination with Galera, making these calls
unnecessary.
EXPLAIN EXTENDED for an UPDATE/DELETE/INSERT/REPLACE statement did not
produce the warning containing the text representation of the query
obtained after the optimization phase. Such warning was produced for
SELECT statements, but not for DML statements.
The patch fixes this defect of EXPLAIN EXTENDED for DML statements.
note that `KILL USER foo` should *not* fail with ER_KILL_DENIED_ERROR
when SHOW PROCESSLIST doesn't show connections of that user.
Because no connections exist or because the caller has no PROCESS -
doesn't matter.
also, fix the error message to make sense
("You are not owner of thread <current connection id>" is ridiculous)
The error string from ER_KILL_QUERY_DENIED_ERROR took a different
type to ER_KILL_DENIED_ERROR for the thread id. This shows
up in differences on 32 big endian arches like powerpc (Deb notation).
Normalize the passing of the THD->id to its real type of my_thread_id,
and cast to (long long) on output. As such normalize the
ER_KILL_QUERY_DENIED_ERROR to that convention too.
Note for upwards merge, convert the type to %lld on new translations
of ER_KILL_QUERY_DENIED_ERROR.
