The server shutdown and start code triggered the valgrind failures
within nptl_pthread_exit_hack_handler on Ubuntu 9.04, x86 (but not amd64)
in rpl_trigger.test file.
For fixing the bug, suppress valgrind failures within nptl_pthread_exit_hack_handler
on Ubuntu 9.04, x86 (but not amd64). Because the server shutdown and start
code has been heavily used in mysql test set.
bzr branch mysql-5.1-performance-version mysql-trunk # Summit
cd mysql-trunk
bzr merge mysql-5.1-innodb_plugin # which is 5.1 + Innodb plugin
bzr rm innobase # remove the builtin
Next step: build, test fixes.
Install procedure does not copy *.inc files located under the mysql-test/t directory.
Therefore, this patch moves the rpl_trigger.inc to the mysql-test/include directory.
The test case fails sporadically on Windows while trying to overwrite an unused
binary log. The problem stems from the fact that MySQL on Windows does not
immediately unlock/release a file while the process that opened and closed it is
still running. In BUG 38603, this issue was circumvented by stopping the MySQL
process, copying the file and then restarting the MySQL process.
Unfortunately, such facilities are not available in the 5.0. Other approaches
such as stopping the slave and issuing change master do not work because the relay
log file and index are not closed when a slave is stopped. So to fix the problem,
we simply don't run on windows the part of the test that was failing.
when used with --tab
1) New syntax: added CHARACTER SET clause to the
SELECT ... INTO OUTFILE (to complement the same clause in
LOAD DATA INFILE).
mysqldump is updated to use this in --tab mode.
2) ESCAPED BY/ENCLOSED BY field parameters are documented as
accepting CHAR argument, however SELECT .. INTO OUTFILE
silently ignored rests of multisymbol arguments.
For the symmetrical behavior with LOAD DATA INFILE the
server has been modified to fail with the same error:
ERROR 42000: Field separator argument is not what is
expected; check the manual
3) Current LOAD DATA INFILE recognizes field/line separators
"as is" without converting from client charset to data
file charset. So, it is supposed, that input file of
LOAD DATA INFILE consists of data in one charset and
separators in other charset. For the compatibility with
that [buggy] behaviour SELECT INTO OUTFILE implementation
has been saved "as is" too, but the new warning message
has been added:
Non-ASCII separator arguments are not fully supported
This message warns on field/line separators that contain
non-ASCII symbols.
If using statement based replication (SBR), repeatedly calling
statements which are unsafe for SBR will cause a warning message
to be written to the error for each statement. This might lead
to filling up the error log and there is no way to disable this
behavior.
The solution is to only log these message (about statements unsafe
for statement based replication) if the log_warnings option is set.
For example:
SET GLOBAL LOG_WARNINGS = 0;
INSERT INTO t1 VALUES(UUID());
SET GLOBAL LOG_WARNINGS = 1;
INSERT INTO t1 VALUES(UUID());
In this case the message will be printed only once:
[Warning] Statement may not be safe to log in statement format.
Statement: INSERT INTO t1 VALUES(UUID())
We disallow the partitioning of a log table. You could however
partition a table first, and then point logging to it. This is
not only against the docs, it also crashes the server.
We catch this case now.
A REPLACE in the MERGE engine is actually a REPLACE
into one (FIRST or LAST) of the underlying MyISAM
tables. So in effect the server works on the meta
data of the MERGE table, while the real insert happens
in the MyISAM table.
The MERGE table has no index, while MyISAM has a
unique index. When a REPLACE into a MERGE table (
and the REPLACE conflicts with a duplicate in a
child table) is done, we try to access the duplicate
key information for the MERGE table. This information
actually does not exist, hence this results in a crash.
The problem can be resolved by modifying the MERGE
engine to provide us the duplicate key information
directly, instead of just returning the MyISAM index
number as the error key. Then the SQL layer (or "the
server") does not try to access the key_info of the
MERGE table, which does not exist.
The current patch modifies the MERGE engine to provide
the position for a record where a unique key violation
occurs.
an assertion in a debug build.
The reason is that the C API doesn't support multiple result sets for prepared
statements and attempting to execute a stored routine which returns multiple result
sets sometimes lead to a network error. The network error sets the diagnostic area
prematurely which later leads to the assert when an attempt is made to set a second
server state.
This patch fixes the issue by changing the scope of the error code returned by
sp_instr_stmt::execute() to include any error which happened during the execution.
To assure that Diagnostic_area::is_sent really mean that the message was sent all
network related functions are checked for return status.
those keywords do nothing in 5.1 (they are meant for future versions, for example featuring the Maria engine)
so they are here removed from the syntax. Adding those keywords to future versions when needed is:
- WL#5034 "Add TRANSACTIONA=0|1 and PAGE_CHECKSUM=0|1 clauses to CREATE TABLE"
- WL#5037 "New ROW_FORMAT value for CREATE TABLE: PAGE"
consumption (CPU) for upgrading a large log table can be intense.
Therefore, truncate the general_log table beforehand if running
mysql_upgrade test with Valgrind.
If the log_bin_trust_function_creators option is not defined, creating a stored
function requires either one of the modifiers DETERMINISTIC, NO SQL, or READS
SQL DATA. Executing a stored function should also follows the same rules if in
STATEMENT mode. However, this was not happening and a wrong error was being
printed out: ER_BINLOG_ROW_RBR_TO_SBR.
The patch makes the creation and execution compatible and prints out the correct
error ER_BINLOG_UNSAFE_ROUTINE when a stored function without one of the modifiers
above is executed in STATEMENT mode.
One of the tests introduced for this bug was failing
because of path size restriction in windows.
Moved the test case to a new test which is disabled under windows.
to wrong result
When using MIXED mode and issuing 'CREATE TEMPORARY TABLE t_tmp',
the statement is logged if the current binlogging mode is
STATEMENT. This causes the slave to replay the instruction and
create the temporary table as well. If there is no switch to ROW
mode, and later on a 'DROP TEMPORARY TABLE t_tmp' is issued, then
this statement will also be logged and the slave will
remove/close the temporary table.
However, if there is a switch to ROW mode between the CREATE and
DROP TEMPORARY table, the DROP statement will not be logged,
leaving the slave with a dangling temporary table.
This patch addresses this, by always logging a DROP TEMPORARY
TABLE IF EXISTS when in mixed mode and a drop statement is issued
for temporary table(s).
mysqld
The problem was that enabling the event scheduler inside a init
file caused the server to crash upon start-up. The crash occurred
because the event scheduler wasn't being initialized before the
commands in the init-file are processed.
The solution is to initialize the event scheduler before the init
file is read. The patch also disables the event scheduler during
bootstrap and makes the bootstrap operation robust in the
presence of background threads.
In create_myisam_from_heap() mark all errors as fatal except
HA_ERR_RECORD_FILE_FULL for a HEAP table.
Not doing so could lead to problems, e.g. in a case when a
temporary MyISAM table gets overrun due to its MAX_ROWS limit
while executing INSERT/REPLACE IGNORE ... SELECT.
The SELECT execution was aborted, but the error was
converted to a warning due to IGNORE clause, so neither 'ok'
nor 'error' packet could be sent back to the client. This
condition led to hanging client when using 5.0 server, or
assertion failure in 5.1.
Problem was that a failing rename just left the partitions at the state
it was at the failure.
Solution was to try to revert the started rename if a failure occured.
> ------------------------------------------------------------
> revno: 2792
> revision-id: sergey.glukhov@sun.com-20090703083500-jq8vhw0tqr37j7te
> parent: bernt.johnsen@sun.com-20090703083610-o7l4s8syz05rc4w0
> committer: Sergey Glukhov <Sergey.Glukhov@sun.com>
> branch nick: mysql-5.0-bugteam
> timestamp: Fri 2009-07-03 13:35:00 +0500
> message:
> Bug#45806 crash when replacing into a view with a join!
> The crash happend because for views which are joins
> we have table_list->table == 0 and
> table_list->table->'any method' call leads to crash.
> The fix is to perform table_list->table->file->extra()
> method for all tables belonging to view.
> ------------------------------------------------------------
> revno: 2772
> revision-id: joro@sun.com-20090615133815-eb007p5793in33p5
> parent: joro@sun.com-20090612140659-4hj1tta9p8wvcw4k
> committer: Georgi Kodinov <joro@sun.com>
> branch nick: B44810-5.0-bugteam
> timestamp: Mon 2009-06-15 16:38:15 +0300
> message:
> Bug #44810: index merge and order by with low sort_buffer_size
> crashes server!
>
> The problem affects the scenario when index merge is followed by a filesort
> and the sort buffer is not big enough for all the sort keys.
> In this case the filesort function will read the data to the end through the
> index merge quick access method (and thus closing the cursor etc),
> but will leave the pointer to the quick select method in place.
> It will then create a temporary file to hold the results of the filesort and
> will add it as a sort output file (in sort.io_cache).
> Note that filesort will copy the original 'sort' structure in an automatic
> variable and restore it after it's done.
> As a result at exiting filesort() we have a sort.io_cache filled in and
> nothing else (as a result of close of the cursors at end of reading data
> through index merge).
> Now create_sort_index() will note that there is a select and will clean it up
> (as it's been used already by filesort() reading the data in). While doing that
> a special case in the index merge destructor will clean up the sort.io_cache,
> assuming it's an output of the index merge method and is not needed anymore.
> As a result the code that tries to read the data back from the filesort output
> will get no data in both memory and disk and will crash.
>
> Fixed similarly to how filesort() does it : by copying the sort.io_cache structure
> to a local variable, removing the pointer to the io_cache (so that it's not freed
> by QUICK_INDEX_MERGE_SELECT::~QUICK_INDEX_MERGE_SELECT) and restoring the original
> structure (together with the valid pointer) after the cleanup is done.
> This is a safe thing to do because all the structures are already cleaned up by
> hitting the end of the index merge's read method (QUICK_INDEX_MERGE_SELECT::get_next())
> and the cleanup code being written in a way that tolerates repeating cleanups.
> ------------------------------------------------------------
> revno: 2763
> revision-id: sergey.glukhov@sun.com-20090602063813-33mh88cz5vpa2jqe
> parent: alexey.kopytov@sun.com-20090601124224-zgt3yov9wou590e9
> committer: Sergey Glukhov <Sergey.Glukhov@sun.com>
> branch nick: mysql-5.0-bugteam
> timestamp: Tue 2009-06-02 11:38:13 +0500
> message:
> Bug#45152 crash with round() function on longtext column in a derived table
> The crash happens due to wrong max_length value which is set on
> Item_func_round::fix_length_and_dec() stage. The value is set to
> args[0]->max_length which is too big in case of LONGTEXT(LONGBLOB) fields.
> The fix is to set max_length using float_length() function.
> ------------------------------------------------------------
> revno: 2733
> revision-id: gshchepa@mysql.com-20090430192037-9p1etcynkglte2j3
> parent: aelkin@mysql.com-20090430143246-zfqaz0t7uoluzdz2
> committer: Gleb Shchepa <gshchepa@mysql.com>
> branch nick: mysql-5.0-bugteam
> timestamp: Fri 2009-05-01 00:20:37 +0500
> message:
> Bug #37362: Crash in do_field_eq
>
> EXPLAIN EXTENDED of nested query containing a error:
>
> 1054 Unknown column '...' in 'field list'
>
> may cause a server crash.
>
>
> Parse error like described above forces a call to
> JOIN::destroy() on malformed subquery.
> That JOIN::destroy function closes and frees temporary
> tables. However, temporary fields of these tables
> may be listed in st_select_lex::group_list of outer
> query, and that st_select_lex may not cleanup them
> properly. So, after the JOIN::destroy call that
> st_select_lex::group_list may have Item_field
> objects with dangling pointers to freed temporary
> table Field objects. That caused a crash.
binlog
The fix for BUG 43929 introduced a regression issue. In a nutshell, when a
statement that changes a non-transactional table fails, it is written to the
binary log with the error code appended. Unfortunately, after BUG 43929, this
failure was flushing the transactional chace causing mismatch between execution
and logging histories. To fix this issue, we avoid flushing the transactional
cache when a commit or rollback is not issued.
When during the optimization an item is moved to the upper select
the item's context left unchanged. This caused wrong result in the
PS/SP mode.
The Item_ident::remove_dependence_processor now sets the context
of the select to which the item is moved to.
