Problem: After the fix for Bug#12589870, a new field that
stores the length of db name was added in the buffer that
stores the query to be executed. Unlike for the plain user
session, the replication execution did not allocate the
necessary chunk in Query-event constructor. This caused an
invalid read while accessing this field.
Solution: We fix this problem by allocating a necessary chunk
in the buffer created in the Query_log_event::Query_log_event()
and store the length of database name.
PROBLEM:
Threads end-up in deadlock due to locks acquired as described
below,
con1: Run Query on a table.
It is important that this SELECT must back-off while
trying to open the t1 and enter into wait_for_condition().
The SELECT then is blocked trying to lock mysys_var->mutex
which is held by con3. The very significant fact here is
that mysys_var->current_mutex will still point to LOCK_open,
even if LOCK_open is no longer held by con1 at this point.
con2: Try dropping table used in con1 or query some table.
It will hold LOCK_open and be blocked trying to lock
kernel_mutex held by con4.
con3: Try killing the query run by con1.
It will hold THD::LOCK_thd_data belonging to con1 while
trying to lock mysys_var->current_mutex belonging to con1.
But current_mutex will point to LOCK_open which is held
by con2.
con4: Get innodb engine status
It will hold kernel_mutex, trying to lock THD::LOCK_thd_data
belonging to con1 which is held by con3.
So while technically only con2, con3 and con4 participate in the
deadlock, con1's mysys_var->current_mutex pointing to LOCK_open
is a vital component of the deadlock.
CYCLE = (THD::LOCK_thd_data -> LOCK_open ->
kernel_mutex -> THD::LOCK_thd_data)
FIX:
LOCK_thd_data has responsibility of protecting,
1) thd->query, thd->query_length
2) VIO
3) thd->mysys_var (used by KILL statement and shutdown)
4) THD during thread delete.
Among above responsibilities, 1), 2)and (3,4) seems to be three
independent group of responsibility. If there is different LOCK
owning responsibility of (3,4), the above mentioned deadlock cycle
can be avoid. This fix introduces LOCK_thd_kill to handle
responsibility (3,4), which eliminates the deadlock issue.
Note: The problem is not found in 5.5. Introduction MDL subsystem
caused metadata locking responsibility to be moved from TDC/TC to
MDL subsystem. Due to this, responsibility of LOCK_open is reduced.
As the use of LOCK_open is removed in open_table() and
mysql_rm_table() the above mentioned CYCLE does not form.
Revision ID for changes,
open_table() = dlenev@mysql.com-20100727133458-m3ua9oslnx8fbbvz
mysql_rm_table() = jon.hauglid@oracle.com-20101116100012-kxep9txz2fxy3nmw
The following scenario crashes our mysql server:
1. set global innodb_file_per_table=1;
2. create table t1(c1 int) engine=innodb;
3. alter table t1 discard tablespace;
4. alter table t1 add unique index(c1);
Step 4 crashes the server. This patch introduces a check on discarded
tablespace to avoid the crash.
rb://1041 approved by Marko Makela
FULLTEXT INDEX AND CONCURRENT DML.
Problem Statement:
------------------
1) Create a table with FT index.
2) Enable concurrent inserts.
3) In multiple threads do below operations repeatedly
a) truncate table
b) insert into table ....
c) select ... match .. against .. non-boolean/boolean mode
After some time we could observe two different assert core dumps
Analysis:
--------
1)assert core dump at key_read_cache():
Two select threads operating in-parallel on same key
root block.
1st select thread block->status is set to BLOCK_ERROR
because the my_pread() in read_block() is returning '0'.
Truncate table made the index file size as 1024 and pread
was asked to get the block of count bytes(1024 bytes)
from offset of 1024 which it cannot read since its
"end of file" and retuning '0' setting
"my_errno= HA_ERR_FILE_TOO_SHORT" and the key_file_length,
key_root[0] is same i.e. 1024. Since block status has BLOCK_ERROR
the 1st select thread enter into the free_block() and will
be under wait on conditional mutex by making status as
BLOCK_REASSIGNED and goes for wait_on_readers(). Other select
thread will also work on the same block and sees the status as
BLOCK_ERROR and enters into free_block(), checks for BLOCK_REASSIGNED
and asserting the server.
2)assert core dump at key_write_cache():
One select thread and One insert thread.
Select thread gets the unlocks the 'keycache->cache_lock',
which allows other threads to continue and gets the pread()
return value as'0'(please see the explanation above) and
tries to get the lock on 'keycache->cache_lock' and waits
there for the lock.
Insert thread requests for the block, block will be assigned
from the hash list and makes the page_status as
'PAGE_WAIT_TO_BE_READ' and goes for the read_block(), waits
in the queue since there are some other threads performing
reads on the same block.
Select thread which was waiting for the 'keycache->cache_lock'
mutex in the read_block() will continue after getting the my_pread()
value as '0' and sets the block status as BLOCK_ERROR and goes to
the free_block() and go to the wait_for_readers().
Now the insert thread will awake and continues. and checks
block->status as not BLOCK_READ and it asserts.
Fix:
---
In the full text code, multiple readers of index file is not guarded.
Hence added below below code in _ft2_search() and walk_and_match().
to lock the key_root I have used below code in _ft2_search()
if (info->s->concurrent_insert)
mysql_rwlock_rdlock(&share->key_root_lock[0]);
and to unlock
if (info->s->concurrent_insert)
mysql_rwlock_unlock(&share->key_root_lock[0]);
INNODB_AUTOINC_LOCK_MODE=1 AND USING TRIGGER
When an insert stmt like "insert into t values (1),(2),(3)" is
executed, the autoincrement values assigned to these three rows are
expected to be contiguous. In the given lock mode
(innodb_autoinc_lock_mode=1), the auto inc lock will be released
before the end of the statement. So to make the autoincrement
contiguous for a given statement, we need to reserve the auto inc
values at the beginning of the statement.
rb://1074 approved by Alexander Nozdrin
dict_table_replace_index_in_foreign_list(): Replace the dropped index
also in the foreign key constraints of child tables that are
referencing this table.
row_ins_check_foreign_constraint(): If the underlying index is
missing, refuse the operation.
rb:1051 approved by Jimmy Yang
BY A CONCURRENT TRANSACTIO
The member function QUICK_RANGE_SELECT::init_ror_merged_scan() performs
a table handler clone. Innodb does not provide a clone operation.
The ha_innobase::clone() is not there. The handler::clone() does not
take care of the ha_innobase->prebuilt->select_lock_type. Because of
this what happens is that for one index we do a locking read, and
for the other index we were doing a non-locking (consistent) read.
The patch introduces ha_innobase::clone() member function.
It is implemented similar to ha_myisam::clone(). It calls the
base class handler::clone() and then does any additional operation
required. I am setting the ha_innobase->prebuilt->select_lock_type
correctly.
rb://1060 approved by Marko
CAUSES RESTORE PROBLEM
Problem Statement:
------------------
mysqldump is not having the dump stmts for general_log and slow_log
tables. That is because of the fix for Bug#26121. Hence, after
dropping the mysql database, and applying the dump by enabling the
logging, "'general_log' table not found" errors are logged into the
server log file.
Analysis:
---------
As part of the fix for Bug#26121, we skipped the dumping of tables
for general_log and slow_log, because the data dump of those tables
are taking LOCKS, which is not allowed for log tables.
Fix:
----
We came up with an approach that instead of taking both meta data
and data dump information for those tables, take only the meta data
dump which doesn't need LOCKS.
As part of fixing the issue we came up with below algorithm.
Design before fix:
1) mysql database is having tables like db, event,... general_log,
... slow_log...
2) Skip general_log and slow_log while preparing the tables list
3) Take the TL_READ lock on tables which are present in the table
list and do 'show create table'.
4) Release the lock.
Design with the fix:
1) mysql database is having tables like db, event,... general_log,
... slow_log...
2) Skip general_log and slow_log while preparing the tables list
3) Explicitly call the 'show create table' for general_log and
slow_log
3) Take the TL_READ lock on tables which are present in the table
list and do 'show create table'.
4) Release the lock.
While taking the meta data dump for general_log and slow_log the
"CREATE TABLE" is replaced with "CREATE TABLE IF NOT EXISTS".
This is because we skipped "DROP TABLE" for those tables,
"DROP TABLE" fails for these tables if logging is enabled.
Customer is applying the dump by enabling logging so, if the dump
has "DROP TABLE" it will fail. Hence, removed the "DROP TABLE"
stmts for those tables.
After the fix we could observe "Table 'mysql.general_log'
doesn't exist" errors initially that is because in the customer
scenario they are dropping the mysql database by enabling the
logging, Hence, those errors are expected. Once we apply the
dump which is taken before the "drop database mysql", the errors
will not be there.
The function mysql_show_binlog_events has a local stack variable
'LOG_INFO linfo;', which is assigned to thd->current_linfo, however
this variable goes out of scope and is destroyed before clean
thd->current_linfo.
The problem is solved by moving 'LOG_INFO linfo;' to function scope.
BUG#11761686 insert_id event is not filtered.
Two issues are covered.
INSERT into autoincrement field which is not the first part in the composed primary key
is unsafe by autoincrement logging design. The case is specific to MyISAM engine
because Innodb does not allow such table definition.
However no warnings and row-format logging in the MIXED mode was done, and
that is fixed.
Int-, Rand-, User-var log-events were not filtered along with their parent
query that made possible them to screw up execution context of the following
query.
Fixed with deferring their execution until the parent query.
******
Bug#11754117
Post review fixes.
Reason:
This is a regression happened because of changes done in code refactoring
in 5.1 from 5.0.
Issue:
While doing "Show tables" lex->verbose was being checked to avoid opening
FRM files to get table type. In case of "Show full table", lex->verbose
is true to indicate table type is required. In 5.0, this check was
present which got missing in >=5.5.
Fix:
Added the required check to avoid opening FRM files unnecessarily in case
of "Show tables".
Currently SHOW MASTER LOGS and SHOW BINARY LOGS require the SUPER
privilege. Monitoring tools (such as MEM) often want to check this
output - for instance MEM generates the SUM of the sizes of the logs
reported here, and puts that in the Replication overview within the MEM
Dashboard.
However, because of the SUPER requirement, these tools often have an
account that holds open the connection whilst monitoring, and can lock
out administrators when the server gets overloaded and reaches
max_connections - there is already another SUPER privileged account
connected, the "monitor".
As SHOW MASTER STATUS, and all other replication related statements,
return with either REPLICATION CLIENT or SUPER privileges, this worklog
is to make SHOW MASTER LOGS and SHOW BINARY LOGS be consistent with this
as well, and allow both of these commands with either SUPER or
REPLICATION CLIENT.
This allows monitoring tools to not require a SUPER privilege any more,
so is safer in overloaded situations, as well as being more secure, as
lighter privileges can be given to users of such tools or scripts.
ORDER BY COUNT(*) LIMIT.
PROBLEM:
With respect to problem in the bug description, we
exhibit different behaviors for the two tables
presented, because innodb statistics (rec_per_key
in this case) are updated for the first table
and not so for the second one. As a result the
query plan gets changed in test_if_skip_sort_order
to use 'index' scan. Hence the difference in the
explain output. (NOTE: We can reproduce the problem
with first table by reducing the number of tuples
and changing the table structure)
The varied output w.r.t the query on the second table
is because of the result in the query plan change.
When a query plan is changed to use 'index' scan,
after the call to test_if_skip_sort_order, we set
keyread to TRUE immedietly. If for some reason
we drop this index scan for a filesort later on,
we fetch only the keys not the entire tuple.
As a result we would see junk values in the result set.
Following is the code flow:
Call test_if_skip_sort_order
-Choose an index to give sorted output
-If this is a covering index, set_keyread to TRUE
-Set the scan to INDEX scan
Call test_if_skip_sort_order second time
-Index is not chosen (note that we do not pass the
actual limit value second time. Hence we do not choose
index scan second time which in itself is a bug fixed
in 5.6 with WL#5558)
-goto filesort
Call filesort
-Create quick range on a different index
-Since keyread is set to TRUE, we fetch only the columns of
the index
-results in the required columns are not fetched
FIX:
Remove the call to set_keyread(TRUE) from
test_if_skip_sort_order. The access function which is
'join_read_first' or 'join_read_last' calls set_keyread anyways.
Background :
In mysql-5.1, in a fix for bug#47485, code has been changed for
mysql client (libmysql/libmysql.c) but corresponding code was not
changed for embedded mysql. In that code change, after execution
of a statement, mysql_stmt_store_result() checks for mysql->state
to be MYSQL_STATUS_STATEMENT_GET_RESULT, instead of
MYSQL_STATUS_GET_RESULT (earlier).
Reason:
In embedded mysql code, after execution, mysql->state was not
set to MYSQL_STATUS_STATEMENT_GET_RESULT, so it was throwing
OUT_OF_SYNC error.
Fix:
Fixed the code in libmysqld/lib_sql.cc to have mysql->state
to be set to MYSQL_STATUS_STATEMENT_GET_RESULT after execution.
Bug#13639204 64111: CRASH ON SELECT SUBQUERY WITH NON UNIQUE INDEX
The crash happened due to wrong calculation
of key length during creation of reference for
sort order index. The problem is that
keyuse->used_tables can have OUTER_REF_TABLE_BIT enabled
but used_tables parameter(create_ref_for_key() func) does
not have it. So key parts which have OUTER_REF_TABLE_BIT
are ommited and it could lead to incorrect key length
calculation(zero key length).
Analysis:
-------------------------------
According to the Manual
(http://dev.mysql.com/doc/refman/5.1/en/identifier-case-sensitivity.html):
"Column, index, stored routine, and event names are not case sensitive on any
platform, nor are column aliases."
In other words, 'lower_case_table_names' does not affect the behaviour of
those identifiers.
On the other hand, trigger names are case sensitive on some platforms,
and case insensitive on others. 'lower_case_table_names' does not affect
the behaviour of trigger names either.
The bug was that SHOW statements did case sensitive comparison
for stored procedure / stored function / event names.
Fix:
Modified the code so that comparison in case insensitive for routines
and events for "SHOW" operation.
As part of this commit, only fixing the test failures due to the actual code fix.
Change the type of purge_sys_t::n_pages_handled and purge_sys_t::handle_limit
to ulonglong from ulint. On a 32 bit system doing ~700 deletes per second the
counters can overflow in ~3.5 months, if they are 32 bit.
Approved by Jimmy Yang over IM.
Analysis:
-------------------------------
According to the Manual
(http://dev.mysql.com/doc/refman/5.1/en/identifier-case-sensitivity.html):
"Column, index, stored routine, and event names are not case sensitive on any
platform, nor are column aliases."
In other words, 'lower_case_table_names' does not affect the behaviour of
those identifiers.
On the other hand, trigger names are case sensitive on some platforms,
and case insensitive on others. 'lower_case_table_names' does not affect
the behaviour of trigger names either.
The bug was that SHOW statements did case sensitive comparison
for stored procedure / stored function / event names.
Fix:
Modified the code so that comparison in case insensitive for routines
and events for "SHOW" operation.
The test case must insert all the records using a single transaction. Otherwise the test
case takes more than 15 minutes and will time out in pb2 and mtr.
FROM BUFFER POOL
rb://975
approved by: Marko Makela
There is a race in lock_validate() where we try to access a page
without ensuring that the tablespace stays valid during the operation
i.e.: it is not deleted. This patch tries to fix that by using an
existing flag (the flag is renamed to make it's name more generic
in line with it's new use).
IN OS_THREAD_EQ
rb://977
approved by: Marko Makela
rw_lock::writer_thread field contains the thread id of current x-holder
or wait-x thread. This field is un-initialized at lock creation and is
written to for the first time when an attempt is made to x-lock.
Current code considers ::writer_thread as valid memory region only when
the lock is held in x-mode (or there is an x-waiter). This is an
overkill and it generates valgrind warnings.
The fix is to consider ::writer_thread as valid memory region once it
has been written to.
Reasoning:
==========
The ::writer_thread can be safely considered valid because:
* We only ever do comparison with current calling threads id.
* We only ever do comparison when ::recursive flag is set
* We always unset ::recursive flag in x-unlock
* Same thread cannot be unlocking and attempting to lock at the same
time
* thread_id recycling is not an issue because before an id is recycled
the thread must leave innodb meaning it must release all locks meaning
it must unset ::recursive flag.
