Problem:-
In case of blob data field, UNION ALL doesn't give correct result.
Analysis:-
In MyISAM table, when we dont want to check for the distinct for particular
key, we set the key_map to zero.
While writing record in MyISAM table, we check the distinct with the help
of keys, by checking whether that key is active in key_map and then writing
the record.
In case of blob field, we are checking for distinct by unique constraint,
where we are not checking whether that unique key is active or not in key_map.
Solution:-
Before checking for distinct, check whether any key is active in key_map.
TO 'MYISAM_SORT_BUFFER_SIZE'
Problem: 'myisam_sort_buffer_size' is a parameter used by
mysqld program only whereas 'sort_buffer_size' is used by
mysqld and myisamchk programs. But the error message printed
when myisamchk program is run with insufficient buffer size
is myisam_sort_buffer_size is too small which may mislead to the
server parameter myisam_sort_buffer_size.
SOLUTION: A parameter 'myisam_sort_buffer_size' is added as an
alias for 'sort_buffer_size' and the 'sort_buffer_size' parameter
is marked as deprecated. So myisamchk also has both the parameters
with the same role.
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]);
CASES RESETS DATA POINTER TO SMAL
ISSUE: Myisamchk doing sort recover
on a table reduces data_file_length.
Maximum size of data file decreases,
lesser number of rows are stored.
SOLUTION: Size of data_file_length is
fixed to the original length.
CASES RESETS DATA POINTER TO SMAL
ISSUE: Myisamchk doing sort recover
on a table reduces data_file_length.
Maximum size of data file decreases,
lesser number of rows are stored.
SOLUTION: Size of data_file_length is
fixed to the original length.
WITH MYISAM_USE_MMAP ENABLED
MySQL server can crash due to segmentation fault when
started with myisam_use_mmap.
The reason behind this being, while making a request to
unmap (munmap) the previously mapped memory (mmap), the
size passed was 7 bytes larger than the size requested at
the time of mapping. This can eventually unmap the adjacent
memory mapped block, belonging to some other memory-map pool.
Hence the subsequent call to mmap can map a region which was
still a valid memory mapped area.
Fixed by removing the extra 7-byte margin which was erroneously
added to the size, used for unmappping.
USING MYISAM_USE_MMAP ON WINDOWS
When OPTIMIZE/REPAIR TABLE is switching to new data file,
old data file is removed while memory mapping is still
active.
With 5.1 implementation of nt_share_delete() it is not
permitted to remove mmaped file.
This fix disables memory mapping for mi_repair() operations.
FULLTEXT INDEXES
myisamchk may create incorrect fulltext index for compressed
tables. Incorrect data pointer size was used while creating
fulltext index.
Problem: in case of wrong data insert into indexed GEOMETRY fields
(e.g. NULL value for a not NULL field) MyISAM reported
"ERROR 126 (HY000): Incorrect key file for table; try to repair it"
due to misuse of the key deletion function.
Fix: always use R-tree key functions for R-tree based indexes
and B-tree key functions for B-tree based indexes.
The LGPL license is used in some legacy code, and to
adhere to current licensing polity, we remove those
files that are no longer used, and reorganize the
remaining LGPL code so it will be GPL licensed from
now on.
Note: This patch only removed LGPL licensed files
in MySQL 5.1, and is the second of a set of
patches to remove LGPL from all trees.
(See Bug# 11840513 for details)
When executing row-ordered-retrieval index merge,
the handler was cloned, but it used the wrong
memory root, so instead of allocating memory
on the thread/query's mem_root, it used the table's
mem_root, resulting in non released memory in the
table object, and was not freed until the table was
closed.
Solution was to ensure that memory used during cloning
of a handler was allocated from the correct memory root.
This was implemented by fixing handler::clone() to also
take a name argument, so it can be used with partitioning.
And in ha_partition only allocate the ha_partition's ref, and
call the original ha_partition partitions clone() and set at cloned
partitions.
Fix of .bzrignore on Windows with VS 2010
attempt to create spatial index on char > 31 bytes".
Attempt to create spatial index on char field with length
greater than 31 byte led to assertion failure on server
compiled with safemutex support.
The problem occurred in mi_create() function which was called
to create a new version of table being altered. This function
failed since it detected an attempt to create a spatial key
on non-binary column and tried to return an error.
On its error path it tried to unlock THR_LOCK_myisam mutex
which has not been not locked at this point. Indeed such an
incorrect behavior was caught by safemutex wrapper and caused
assertion failure.
This patch fixes the problem by ensuring that mi_create()
doesn't releases THR_LOCK_myisam mutex on error path if it was
not acquired.
DROP/CREATE SCHEMA, CREATE TABLE, REPAIR.
The cause of assert was concurrent execution of
DROP DATABASE and REPAIR TABLE where first statement
deleted table's file .TMD at the same time as
REPAIR TABLE tried to read file details from the old file
that was just removed.
Additionally was fixed trouble when DROP TABLE try delete
all files belong to table being dropped at the same time
when REPAIR TABLE statement has just deleted .TMD file.
No regression test added because this would require adding a
sync point to mysys/my_redel.c. Since this bug is not present in
5.5+, adding test coverage was considered unnecessary.
The patch has been verified using RQG testing.
Fix assorted warnings that are generated in optimized builds.
Most of it is silencing variables that are set but unused.
This patch also introduces the MY_ASSERT_UNREACHABLE macro
which helps the compiler to deduce that a certain piece of
code is unreachable.
Use UNINIT_VAR workaround instead of LINT_INIT. The former can
also be used to silence false-positives in non-debug builds as
it actually does not cause new code to be generated.
Fix warnings flagged by the new warning option -Wunused-but-set-variable
that was added to GCC 4.6 and that is enabled by -Wunused and -Wall. The
option causes a warning whenever a local variable is assigned to but is
later unused. It also warns about meaningless pointer dereferences.
Apart strict-aliasing warnings, fix the remaining warnings
generated by GCC 4.4.4 -Wall and -Wextra flags.
One major source of warnings was the in-house function my_bcmp
which (unconventionally) took pointers to unsigned characters
as the byte sequences to be compared. Since my_bcmp and bcmp
are deprecated functions whose only difference with memcmp is
the return value, every use of the function is replaced with
memcmp as the special return value wasn't actually being used
by any caller.
There were also various other warnings, mostly due to type
mismatches, missing return values, missing prototypes, dead
code (unreachable) and ignored return values.
strict aliasing violations.
Essentially, the problem is that large parts of the server were
developed in simpler times (last decades, pre C99 standard) when
strict aliasing and compilers supporting such optimizations were
rare to non-existent. Thus, when compiling the server with a modern
compiler that uses strict aliasing rules to perform optimizations,
there are several places in the code that might trigger undefined
behavior.
As evinced by some recent bugs, GCC does a somewhat good of job
misoptimizing such code, but on the other hand also gives warnings
about suspicious code. One problem is that the warnings aren't
always accurate, yet we can't afford to just shut them off as we
might miss real cases. False-positive cases are aggravated mostly
by casts that are likely to trigger undefined behavior.
The solution is to start a cleanup process focused on fixing and
reducing the amount of strict-aliasing related warnings produced
by GCC and others compilers. A good deal of noise reduction can
be achieved by just removing useless casts that are product of
historical cruft and are likely to trigger undefined behavior if
dereferenced.
data and index files
It was possible if DATA/INDEX DIRECTORY is pointing to
symlinked MySQL data home directory.
Do not allow to drop data/index files implicitly symlinked
to data home directory. For such tables remove symlink only.
Repairing MyISAM table with fulltext indexes and low
myisam_sort_buffer_size may crash the server.
Estimation of number of index entries was done incorrectly,
causing further assertion failure or server crash.
Docs note: min value for myisam_sort_buffer_size has been
changed from 4 to 4096.
Invalid memory read if HANDLER ... READ NEXT is executed
after failed (e.g. empty table) HANDLER ... READ FIRST.
The problem was that we attempted to perform READ NEXT,
whereas there is no pivot available from failed READ FIRST.
With this fix READ NEXT after failed READ FIRST equals
to READ FIRST.
This bug affects MyISAM tables only.
When MyISAM writes newly created index page it may be
initialized partially. In other words some bytes of
sensible data and uninitialized tail of the page may
go into index file.
Under certain rare circumstances these hunks of memory
may contain data that would be otherwise inaccessible
to user, like passwords or data from other tables.
Fixed by initializing memory for temporary MyISAM key
buffer to '\0'.
No test case for this fix as it is heavily covered by
existing tests.
myisam tables
Queries following TRUNCATE of partitioned MyISAM table
may crash server if myisam_use_mmap is true.
Internally this is MyISAM bug, but limited to partitioned
tables, because MyISAM doesn't use ::delete_all_rows()
method for TRUNCATE, but goes via table recreate instead.
MyISAM didn't properly fall back to non-mmaped I/O after
mmap() failure. Was not repeatable on linux before, likely
because (quote from man mmap):
SUSv3 specifies that mmap() should fail if length is 0.
However, in kernels before 2.6.12, mmap() succeeded in
this case: no mapping was created and the call returned
addr. Since kernel 2.6.12, mmap() fails with the error
EINVAL for this case.
index cardinalities=1
Parallel repair didn't poroperly update index cardinality
in certain cases.
When myisam_sort_buffer_size is not enough to store all
keys, index cardinality was updated before index was
actually written, when no index statistic is available.
Spatial indexes were not checking for out-of-record condition in
the handler next command when the previous command didn't found
rows.
Fixed by making the rtree index to check for end of rows condition
before re-using the key from the previous search.
Fixed another crash if the tree has changed since the last search.
Added a test case for the other error.
insert...select
Queries following bulk insert into an empty MyISAM table
may break it. This was pure MyISAM problem.
When bulk insert into an empty table is complete, MyISAM
may want to enable indexes via repair by sort. If repair
by sort fails (e.g. insufficient buffer), MyISAM failover
to repair with key cache, requesting repair of data file.
Repair of data file performs data file substitution. This
means that current table instance will point to new data
file. Other cached table instances are still pointing to
an old, deleted data file.
This is fixed by not requesting repair of data file
during enable indexes.
Explicit REPAIR is not affected, since it flushes all
table instances.
Bulk REPLACE or bulk INSERT ... ON DUPLICATE KEY UPDATE may
break dynamic record MyISAM table.
The problem is limited to bulk REPLACE and INSERT ... ON
DUPLICATE KEY UPDATE, because only these operations may
be done via UPDATE internally and may request write cache.
When flushing write cache, MyISAM may write remaining
cached data at wrong position. Fixed by requesting write
cache to seek to a correct position.
Performing fulltext prefix search (a word with truncation
operator) may cause a dead-loop. ft_min_word_len value
doesn't matter actually.
The problem was introduced along with "smarter index merge"
optimization.
