The reason for the bug was that replaying of a query on slave could not be possible since its event
was recorded with the killed error. Due to the specific of handling INSERT, which per-row-while-loop is
unbreakable to killing, the query on transactional table should have not appeared in binlog unless
there was a call to a stored routine that got interrupted with killing (and then there must be an error
returned out of the loop).
The offered solution added the following rule for binlogging of INSERT that accounts the above
specifics:
For INSERT on transactional-table if the error was not set the only raised flag
is harmless and is ignored via masking out on time of creation of binlog event.
For both table types the combination of raised error and KILLED flag indicates that there
was potentially partial execution on master and consistency is under the question.
In that case the code continues to binlog an event with an appropriate killed error.
The fix relies on the specified behaviour of stored routine that must propagate the error
to the top level query handling if the thd->killed flag was raised in the routine execution.
The patch adds an arg with the default killed-status-unset value to Query_log_event::Query_log_event.
sql/log_event.cc:
killed_status as the value of thd->killed can be passed as an arg to the constructor.
if the value is different from the default the arg is set to the current thd->killed value.
A caller might need to masquerade thd->killed with THD::NOT_KILLED.
So far only mysql_insert() uses such explicit way to tell the constructor about killing status.
sql/log_event.h:
default arg to the constructor with meaning of killed status of the query.
if the arg is not explicitly provided the status of thd->killed will be snapshot
inside of the constuctor, which is potentially incorrect (see bug#27571)
sql/sql_class.h:
extending killed_state with no-state member.
sql/sql_insert.cc:
ignore the KILLED flag incl KILL_BAD_DATA when the INSERT query event
is created without an `error';
sql/sql_update.cc:
Suggestion how to fix bug#27571 as comments.
mysql-test/r/binlog_killed.result:
new result file
mysql-test/t/binlog_killed.test:
regression tests also apply for bug27563, BUG#27565
into whalegate.ndb.mysql.com:/home/tomas/mysql-5.0-ndb
ndb/src/kernel/blocks/dbdih/DbdihMain.cpp:
Auto merged
ndb/src/kernel/blocks/ndbcntr/NdbcntrMain.cpp:
Auto merged
ndb/test/run-test/daily-basic-tests.txt:
Auto merged
ndb/src/kernel/blocks/ERROR_codes.txt:
manual merge
ndb/test/ndbapi/testNodeRestart.cpp:
manual merge
so that othernodes dont get confused after having recevied status from master
and then tries to update it self
ndb/src/kernel/blocks/ERROR_codes.txt:
error 1001, delay node_failrep
ndb/src/kernel/blocks/ndbcntr/NdbcntrMain.cpp:
error 1001, delay node_failrep
ndb/test/ndbapi/testNodeRestart.cpp:
testcase
ndb/test/run-test/daily-basic-tests.txt:
testcase
- A race condition caused brief unavailablility when trying to acccess
a table.
- The unprotected variable 'grant_option' wasn't intended to alternate
during normal execution. Variable initialization moved to grant_init
a lines responsible for the alternation are removed.
sql/mysqld.cc:
- Added documentation to describe grant_option flag.
sql/sql_acl.cc:
- This patch removes lines which causes grant_option to alternate and moves
variable initialization to the grant_init()-function.
include/m_string.h:
Reduced the number of elements in log_10[] and log_01[] to not exceed DBL_MAX.
sql/field.cc:
Avoid the warning on Windows.
strings/strtod.c:
Reduced the number of elements in log_10[] and log_01[] to not exceed DBL_MAX.
- check multi update as well as update
- this bug is not present in 5.0, but execution patch is wrong, so there are probably other bugs
mysql-test/r/ndb_basic.result:
Bug #28719: multi pk update ignore corrupts data
- add test + backport some tests from 5.1
mysql-test/t/ndb_basic.test:
Bug #28719: multi pk update ignore corrupts data
- add test + backport some tests from 5.1
into polly.local:/home/kaa/src/maint/mysql-5.0-maint
sql/field.cc:
Auto merged
sql/field.h:
Auto merged
sql/item_cmpfunc.cc:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/mysqld.cc:
Auto merged
strings/strtod.c:
Auto merged
When storing a large number to a FLOAT or DOUBLE field with fixed length, it could be incorrectly truncated if the field's length was greater than 31.
This patch also does some code cleanups to be able to reuse code which is common between Field_float::store() and Field_double::store().
include/m_string.h:
Added declarations for log_10 and log_01 from strtod.c
mysql-test/r/type_float.result:
Added the testcase for bug #28121 "INSERT or UPDATE into DOUBLE(200,0) field being truncated to 31 digits"
mysql-test/t/type_float.test:
Added the testcase for bug #28121 "INSERT or UPDATE into DOUBLE(200,0) field being truncated to 31 digits"
sql/field.cc:
Moved common code from Field_float::store() and Field_double:store() to Field_real::truncate()
Fixed the algorithm to not truncate large input numbers if the field length is greater than 31.
Fixed rounding to not depend on FLT_MAX/DBL_MAX constants.
sql/field.h:
Moved not_fixed member from Field_double to Field_real to allow code reuse between Field_float::store() and Field_double::store()
Added truncate() method to Field_real which is used by both Field_float and Field_double
sql/init.cc:
log_10[] and log_01[] are now defined as statical arrays in strtod.c, no need to pre-computed them.
sql/item_cmpfunc.cc:
log_01[] now starts from 1e0, not from 1e-1 for consistency.
sql/mysql_priv.h:
Moved log_10[] and log_01[] from mysqld.cc to libmystrings.
sql/mysqld.cc:
Moved log_10[] and log_01[] from mysqld.cc to libmystrings.
strings/strtod.c:
Define and use log_10[] and log_01[] as static arrays of constants instead of values pre-computed at startup.
into mysql.com:/nfsdisk1/lars/MERGE/mysql-5.0-merge
mysql-test/mysql-test-run.pl:
Auto merged
mysql-test/t/disabled.def:
Auto merged
sql/sql_load.cc:
Auto merged
into bk-internal.mysql.com:/data0/bk/mysql-5.0-opt
client/mysqldump.c:
Auto merged
mysql-test/r/mysqldump.result:
Auto merged
mysql-test/r/ps_2myisam.result:
Auto merged
mysql-test/r/ps_3innodb.result:
Auto merged
mysql-test/r/ps_4heap.result:
Auto merged
mysql-test/r/ps_5merge.result:
Auto merged
mysql-test/r/ps_6bdb.result:
Auto merged
mysql-test/r/ps_7ndb.result:
Auto merged
sql/field.h:
Auto merged
sql/item_func.cc:
Auto merged
sql/my_decimal.h:
Auto merged
sql/sql_base.cc:
Auto merged
fixed by using flush logs, dumping, then doing PURGE MASTER LOGS TO 'binfile', instead
of deleting the log files at the beginning.
NOTE: previously the delete-master-logs would reset the log names back to
filename.00001. Now the trailing number doesn't get reset. This may need to be
documented.
client/mysqldump.c:
changed the code the --delete-master-logs option is used from this:
take locks
delete bin logs
do dump (if this is a lock-based dump)
release locks
do dump (if this is a consistent-read-dump)
to this:
take locks
flush logs
remember the name of the new log
do dump (if this is a lock-based dump)
release locks
do dump (if this is a consistent-read-dump)
fflush output sql file if specified, to ensure the backup is commited to disk
--- yes, dump succeeded ---
do PURGE MASTER LOGS TO up to the new log
constant outer tables did not return null complemented
rows when conditions were evaluated to FALSE.
Wrong results were returned because the conditions over constant
outer tables, when being pushed down, were erroneously enclosed
into the guard function used for WHERE conditions.
mysql-test/r/join_outer.result:
Added a test case for bug #28571.
mysql-test/t/join_outer.test:
Added a test case for bug #28571.
sql/sql_select.cc:
Fixed bug #28571. Outer join queries with ON conditions over
constant outer tables did not return null complemented
rows when conditions were evaluated to FALSE.
Wrong results were returned because the conditions over constant
outer tables, when being pushed down, were erroneously enclosed
into the guard function used for WHERE conditions.
The problem is fixed in the function make_join_select. Now the
conditions over constant tables from ON expressions are pushed
down after the conditions from WHERE has been pushed down.
The root cause of this bug is related to the function skip_rear_comments,
in sql_lex.cc
Recent code changes in skip_rear_comments changed the prototype from
"const uchar*" to "const char*", which had an unforseen impact on this test:
(endp[-1] < ' ')
With unsigned characters, this code filters bytes of value [0x00 - 0x20]
With *signed* characters, this also filters bytes of value [0x80 - 0xFF].
This caused the regression reported, considering cyrillic characters in the
parameter name to be whitespace, and truncated.
Note that the regression is present both in 5.0 and 5.1.
With this fix:
- [0x80 - 0xFF] bytes are no longer considered whitespace.
This alone fixes the regression.
In addition, filtering [0x00 - 0x20] was found bogus and abusive,
so that the code now filters uses my_isspace when looking for whitespace.
Note that this fix is only addressing the regression affecting UTF-8
in general, but does not address a more fundamental problem with
skip_rear_comments: parsing a string *backwards*, starting at end[-1],
is not safe with multi-bytes characters, so that end[-1] can confuse the
last byte of a multi-byte characters with a characters to filter out.
The only known impact of this remaining issue affects objects that have to
meet all the conditions below:
- the object is a FUNCTION / PROCEDURE / TRIGGER / EVENT / VIEW
- the body consist of only *1* instruction, and does *not* contain a
BEGIN-END block
- the instruction ends, lexically, with <ident> <whitespace>* ';'?
For example, "select <ident>;" or "return <ident>;"
- The last character of <ident> is a multi-byte character
- the last byte of this character is ';' '*', '/' or whitespace
In this case, the body of the object will be truncated after parsing,
and stored in an invalid format.
This last issue has not been fixed in this patch, since the real fix
will be implemented by Bug 25411 (trigger code truncated), which is caused
by the very same code.
The real problem is that the function skip_rear_comments is only a
work-around, and should be removed entirely: see the proposed patch for
bug 25411 for details.
sql/sp_head.cc:
In skip_rear_comments,
Filter out only whitespace, not other (non ascii or control) valid characters
sql/sql_lex.cc:
In skip_rear_comments,
Filter out only whitespace, not other (non ascii or control) valid characters
sql/sql_lex.h:
In skip_rear_comments,
Filter out only whitespace, not other (non ascii or control) valid characters
sql/sql_view.cc:
In skip_rear_comments,
Filter out only whitespace, not other (non ascii or control) valid characters
tests/mysql_client_test.c:
Bug#27876 (SF with cyrillic variable name fails during execution (regression))
On some Linux distributions with both LinuxThreads and NPTL glibc versions available, statically built binaries can crash, because linker defaults to LinuxThreads when linking statically, but calls to external libraries (like libnss) are resolved to NPTL versions.
Since there is nothing we can do in the code to work that around, just give user an advice on how to fix that, if a crash happened on such a binary/OS combination.
sql/mysqld.cc:
Advise a user on possible solutions in case of a crash in a statically built binary on a system with both LinuxThreads and NPTL available.
- Add checks to make sure net has a vio assigned
- For example bootstrap will create a fake "net" with vio
set to 0
sql/net_serv.cc:
Add checks to make sure net has a vio assigned
sometimes `mysqldump --hex-blob' overruned output buffer by '\0' byte.
The dump_table() function has been fixed to reserve 1 byte more for the
last '\0' byte of dumped string.
client/mysqldump.c:
Fixed bug #28522.
The dump_table() function has been fixed to reserve 1 byte more for the
last '\0' byte of dumped string.
mysql-test/t/mysqldump.test:
Updated test case for bug #28522.
mysql-test/r/mysqldump.result:
Updated test case for bug #28522.
- Make the two "my_net_set*" function only visible when included
from my_global.h
include/mysql_com.h:
Make the two "my_net_set*" functions only
visible when included from my_global.h
into pilot.blaudden:/home/msvensson/mysql/mysql-5.0-maint
libmysql/libmysql.c:
Auto merged
sql-common/client.c:
Auto merged
sql/mysql_priv.h:
Auto merged
sql/net_serv.cc:
Auto merged
sql/set_var.cc:
Auto merged
sql/sql_parse.cc:
Auto merged
sql/sql_repl.cc:
Auto merged
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
Problem:
binlogging PS' we may produce syntacticly incorrect queries in the binlog replacing
some parameters with variable names (instead of variable values).
E.g. in the reported case of "limit ?" clause: replacing "?" with "@var"
produces "limit @var" which is not a correct SQL syntax.
Also it may lead to different query execution on slave if we
set and use a variable in the same statement, e.g.
"insert into t1 values (@x:=@x+1, ?)"
Fix: make the stored statement string created upon its execution use variable values
(instead of names) to fill placeholders.
mysql-test/r/ctype_cp932_binlog.result:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- result adjusted.
mysql-test/r/ctype_cp932_notembedded.result:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- result adjusted.
mysql-test/r/rpl_user_variables.result:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- test result.
mysql-test/t/ctype_cp932_binlog.test:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- test adjusted.
mysql-test/t/rpl_user_variables.test:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- test case.
sql/sql_prepare.cc:
Fix for
bug #26842: master binary log contains invalid queries - replication fails
bug #12826: Possible to get inconsistent slave using SQL syntax Prepared Statements
- set val to the variable's value (escaped if needed) then insert it into the query
string in the position of the placeholder. We don't need to call
get_var_with_binlog() here as there is no trace of the variable's name in the binlog.
- The "mysql client in mysqld"(which is used by
replication and federated) should use alarms instead of setting
socket timeout value if the rest of the server uses alarm. By
always calling 'my_net_set_write_timeout'
or 'my_net_set_read_timeout' when changing the timeout value(s), the
selection whether to use alarms or timeouts will be handled by
ifdef's in those two functions.
- Move declaration of 'vio_timeout' into "vio_priv.h"
include/mysql_com.h:
Move the net_set_*_timeout function declarations to mysql_com.h and
rename to my_net_set_*_timeout to avoid name clashes
include/violite.h:
Move declaration of 'vio_timeout' to vio_priv.h (to make
the function as private as possible)
libmysql/libmysql.c:
Use my_net_read_timeout or my_net_write_timeout when setting
the timeouts. Move the global variables for my_net_read/my_write_timeout
into the only place where they are used. Thus removing them...
server-tools/instance-manager/mysql_connection.cc:
Use my_net_read_timeout or my_net_write_timeout when setting the timeouts
sql-common/client.c:
Use my_net_read_timeout or my_net_write_timeout when setting the timeouts
sql/mysql_priv.h:
Move the net_set_*_timeout function declarations to mysql_com.h
sql/net_serv.cc:
No need to cast the net->write_timeout value from "uint" to "uint"
sql/set_var.cc:
Rename net_set_*_timeout to my_net_set_*_timeout
sql/sql_client.cc:
Use my_net_read_timeout or my_net_write_timeout when setting the timeouts
sql/sql_parse.cc:
Rename net_set_*_timeout to my_net_set_*_timeout
sql/sql_repl.cc:
Rename net_set_*_timeout to my_net_set_*_timeout
vio/vio_priv.h:
Move declaration of 'vio_timeout' to vio_priv.h
vio/viosocket.c:
Cleanup 'vio_timeout'
- Use "const void*" on POSIX and "const char*" on windows for
setsockopt
- Add DBUG_PRINT's
- Add comment about why we don't have an implementation of
vio_timeout for platforms not supporting SO_SNDTIMEO or SO_RCVTIMEO
into pippilotta.erinye.com:/shared/home/df/mysql/build/mysql-5.0-build
configure.in:
Auto merged
include/my_global.h:
Auto merged
sql/item_func.cc:
Auto merged
strings/strtod.c:
Auto merged
CHECK OPTION and a subquery in WHERE condition.
The abort was triggered by setting the value of join->tables for
subqueries in the function JOIN::cleanup. This function was called
after an invocation of the JOIN::join_free method for subqueries
used in WHERE condition.
mysql-test/r/view.result:
Added a test case for bug #28561.
mysql-test/t/view.test:
Added a test case for bug #28561.
sql/sql_select.cc:
Fixed bug #28561: assertion abort for update on multi-table view with
CHECK OPTION and a subquery in WHERE condition.
The abort was triggered by setting the value of join->tables for
subqueries in the function JOIN::cleanup. This function was called
after an invocation of the JOIN::join_free method for subqueries
used in WHERE condition.
Setting the value of join->tables to for a subquery created serious
problems for checking WHERE condition after update of the multi-table
view as this check is performed in the do_select function right
after a call of the JOIN::join_free method.
In fact setting join->tables to 0 in JOIN::cleanup is not needed
anywhere in the current code.
