MySQL Community Server 8.0.20 has been released (part 2/2)
Posted by: Bjørn Munch
Date: April 27, 2020 12:57PM
Date: April 27, 2020 12:57PM
[ This is part 2 of the announcement ]
Bugs Fixed
* Performance: Certain queries against tables with spatial
indexes were not performed as efficiently following an
upgrade from MySQL 5.7 to MySQL 8.0. (Bug #94655, Bug
#29488350)
References: See also: Bug #89551, Bug #27499984.
* NDB Cluster: NDB defines one SPJ worker per node owning a
primary partition of the root table. If this table used
read from any replica, DBTC put all SPJ workers in the
same DBSPJ instance, which effectively removed the use of
some SPJ workers. (Bug #30639165)
* NDB Cluster: Executing the SHOW command using an ndb_mgm
client binary from NDB 8.0.16 or earlier to access a
management node running NDB 8.0.17 or later produced the
error message Unknown field: is_single_user. (Bug
#30599413)
References: See also: Bug #16275500.
* InnoDB: A CREATE UNDO TABLESPACE operation that specified
an undo data file name without specifying a path removed
an existing undo data file of the same name from the
directory specified by innodb_undo_directory variable.
The file name conflict check was performed on the data
directory instead of the directory specified by the
innodb_undo_directory variable. (Bug #30908328, Bug
#98628)
* InnoDB: In debug builds, a regression introduced in MySQL
8.0.19 slowed down mutex and rw-lock deadlock debug
checks. (Bug #30886393)
References: This issue is a regression of: Bug #30628872.
* InnoDB: Valgrind testing raised an error indicating that
a conditional jump or move depends on an uninitialized
value. The error was a false-positive due to invalid
validation logic. (Bug #30837136)
* InnoDB: Missing barriers in rw_lock_debug_mutex_enter()
(in source file sync0debug.cc) could cause a thread to
wait without ever being woken up. (Bug #30819167)
* InnoDB: To improve server initialization speed,
fallocate() is now used to allocate space for redo log
files. (Bug #30804431)
* InnoDB: A data dictionary table open function was
implemented with incorrect lock ordering. (Bug #30782103,
Bug #97825)
* InnoDB: Changes to parallel read threads functionality
introduced in MySQL 8.0.17 caused a degradation in SELECT
COUNT(*) performance. Pages were read from disk
unnecessarily. (Bug #30766089)
* InnoDB: DDL logging was not performed for SQL operations
executed by the bootstrap thread using the init_file
startup variable, causing files to be left behind that
should have been removed during a post-DDL stage. (Bug
#30721214)
* InnoDB: Adding an index on a column cast as a JSON array
on a table with a specific number of records failed with
an "Incorrect key file for table" error. (Bug #30709525,
Bug #98098)
* InnoDB: A Valgrind error reported that an uninitialized
lock->writer_thread value was used in a conditional jump.
(Bug #30694177)
* InnoDB: An internal buffer pool statistics counter
(n_page_gets) was partitioned by page number to avoid
contention when accessed by multiple threads. (Bug
#30604841, Bug #97822)
* InnoDB: A tablespace import operation failed with a
schema mismatch error due to the .cfg file and the data
dictionary both containing default values for a column
that was added using ALGORITHM=INSTANT. An error should
only occur if default values differ. (Bug #30561144)
* InnoDB: A slow shutdown failed to flush some GTIDs,
requiring recovery of unflushed GTIDs from the undo log.
(Bug #30548229)
* InnoDB: A broken alignment requirement in the code that
allocates a prefix in memory for Performance Schema
memory allocations caused a failure on MySQL builds
optimized for macOS and FreeBSD. (Bug #30530857)
* InnoDB: Adding a virtual column raised an assertion
failure due to data that was missing from the new data
dictionary object created for the table. (Bug #30524263)
* InnoDB: A required latch was not taken when checking the
mode of an undo tablespace. A required latch was also not
taken when checking whether an undo tablespace is empty.
(Bug #30509134)
* InnoDB: Allocating an update undo log segment to an XA
transaction for persisting a GTID value before the
transaction performed any data modifications caused a
failure. (Bug #30456328)
* InnoDB: A query executed on a partitioned table with a
discarded tablespace raised an assertion failure. (Bug
#30437407, Bug #97271)
* InnoDB: The row_upd_clust_rec_by_insert function, which
marks a clustered index record as deleted and inserts an
updated version of the record into the clustered index,
passed an incorrect n_ext value (the total number of
external fields) to lower level functions, causing an
assertion failure. (Bug #30437378)
* InnoDB: During a cloning operation, writes to the data
dictionary buffer table at shutdown were too late,
causing a failure. Newly generated dirty pages were not
being flushed. (Bug #30427369, Bug #30405535, Bug
#30405535)
* InnoDB: An operation performed with the
innodb_buffer_pool_evict debug variable set to
uncompressed caused an assertion failure. (Bug #30405531)
* InnoDB: Read-write lock code (rw_lock_t) that controls
ordering of access to the boolean recursive flag and the
writer thread ID using GCC builtins or os_mutex when the
builtins are not available, was revised to use C++
std::atomic in some instances.
Thanks to Yibo Cai from ARM for the contribution. (Bug
#30401416, Bug #97150)
* InnoDB: A failure occurred while upgrading from MySQL 5.7
to MySQL 8.0. A server data dictionary object was missing
information about the FTS_DOC_ID column and
FTS_DOC_ID_INDEX that remain after dropping a FULLTEXT
index. (Bug #30357954)
* InnoDB: Unnecessary messages about parallel scans were
printed to the error log. (Bug #30330448)
* InnoDB: During upgrade from MySQL 5.7 to MySQL 8.0,
clustered indexes named GEN_CLUST_INDEX are renamed to
PRIMARY, which resulted in duplicate entries for the
clustered indexes being added to the
mysql.innodb_index_stats table. (Bug #30330448)
* InnoDB: Various internal functions computed write event
slots in an inconsistent manner. (Bug #30228108, Bug
#96519)
* InnoDB: Under specific circumstances, it was possible
that tablespace encryption key information would not be
applied during the redo log apply phase of crash
recovery. (Bug #30209760)
* InnoDB: A file operation failure caused the page tracking
archiver to fail, which in turn caused the main thread to
hang, resulting in an assertion failure. Also,
incorrectly, the page tracking archiver remained enabled
in innodb_read_only mode. (Bug #30202643)
* InnoDB: An index corruption error was reported when
attempting to import a tablespace containing a table
column that was added using ALGORITHM=INSTANT. The error
was due to missing metadata associated with the instantly
added column. (Bug #30191523, Bug #96477)
* InnoDB: A transaction attempting to fetch an LOB record
encountered a null LOB reference, causing an assertion
failure. However, the null LOB reference was valid in
this particular scenario because the LOB value was not
yet fully written. (Bug #30144303)
* InnoDB: During a parallel read operation, the rollback of
a table load operation while autocommit was disabled
resulted in a server to exit due to assertion code that
did not account for the possibility of tree structure
changes during a parallel read. (Bug #30060690)
* InnoDB: The current size value maintained in a rollback
segment memory object was found to be invalid, causing an
assertion failure in function trx_purge_free_segment(). A
validation routine (trx_rseg_t::validateCurrSize()) was
added to verify the current size value. (Bug #29947027)
* InnoDB: A prepared statement executed with invalid
parameter values raised an assertion failure. (Bug
#29880907)
* InnoDB: An add column operation caused an assertion
failure. The failure was due to a dangling pointer. (Bug
#29866408)
References: This issue is a regression of: Bug #28491099.
* InnoDB: Updating certain InnoDB system variables that
take string values raised invalid read errors during
Valgrind testing. (Bug #29717909, Bug #95215)
* InnoDB: Redo log records for modifications to undo
tablespaces increased in size in MySQL 8.0 due to a
change in undo tablespace ID values, which required
additional bytes. The change in redo log record size
caused a performance regression in workloads with heavy
write I/O. To address this issue, the redo log format was
modified to reduce redo log record size for modifications
to undo tablespaces. (Bug #29536710)
* InnoDB: Additional information about InnoDB file writes,
including progress data, is now printed to the error log.
(Bug #29472295, Bug #94634)
* InnoDB: An insert statement on a table with a spatial
index raised a record type mismatch assertion due to a
tuple corruption. (Bug #29465567)
* InnoDB: A function that calculates undo log record size
could calculate an incorrect length value in the case of
a corrupted undo log record, resulting in a malloc
failure. Assertion code was added to detect incorrect
calculations. (Bug #29448406, Bug #82734)
* Replication: The thread used by Group Replication's
message service was not correctly registered by the
Performance Schema instrumentation, so the thread actions
were not visible in Performance Schema tables. (Bug
#30824676)
* Replication: Group Replication initiates and manages
cloning operations for distributed recovery, but group
members that have been set up to support cloning may also
participate in cloning operations that a user initiates
manually. In releases before MySQL 8.0.20, you could not
initiate a cloning operation manually if the operation
involved a group member on which Group Replication was
running. From MySQL 8.0.20, you can do this, provided
that the cloning operation does not remove and replace
the data on the recipient. The statement to initiate the
cloning operation must therefore include the DATA
DIRECTORY clause if Group Replication is running. (Bug
#30798640)
* Replication: For Group Replication channels, issuing the
CHANGE MASTER TO statement with the PRIVILEGE_CHECKS_USER
option while Group Replication was running caused the
channel's relay log files to be deleted. Transactions
that had been received and queued in the relay log, but
not yet applied, could be lost in this situation. The
CHANGE MASTER TO statement can now only be issued when
Group Replication is not running. (Bug #30655369)
* Replication: Group Replication's failure detection
mechanism raises a suspicion if a server stops sending
messages, and the member is eventually expelled provided
that a majority of the group members are still
communicating. However, the failure detection mechanism
did not take into account the situation where one or more
of the group members in the majority had actually already
been marked for expulsion, but had not yet been removed
from the group. Where the network was unstable and
members frequently lost and regained connection to each
other in different combinations, it was possible for a
group to end up marking all its members for expulsion,
after which the group would cease to exist and have to be
set up again.
Group Replication's Group Communication System (GCS) now
tracks the group members that have been marked for
expulsion, and treats them as if they were in the group
of suspect members when deciding if there is a majority.
This ensures at least one member remains in the group and
the group can continue to exist. When an expelled member
has actually been removed from the group, GCS removes its
record of having marked the member for expulsion, so that
the member can rejoin the group if it is able to. (Bug
#30640544)
* Replication: While an SQL statement was in the process of
being rewritten for the binary log so that sensitive
information did not appear in plain text, if a SHOW
PROCESSLIST statement was used to inspect the query, the
query could become corrupted when it was written to the
binary log, causing replication to stop. The process of
rewriting the query is now kept private, and the query
thread is updated only when rewriting is complete. (Bug
#30569003, Bug #97531, Bug #30654405)
* Replication: When a GRANT or REVOKE statement is only
partially executed, an incident event is logged in the
binary log, which makes the replication slave's applier
thread stop so that the slave can be reconciled manually
with the master. Previously, if a failed GRANT or REVOKE
statement was the first statement executed in the
session, no GTID was applied to the incident event
(because the cache manager did not yet exist for the
session), causing an error on the replication slave.
Also, no incident event was logged in the situation where
a GRANT statement created a user but then failed because
the privileges had been specified incorrectly, again
causing an error on the replication slave. Both these
issues have now been fixed. (Bug #30566518, Bug
#30324661)
* Replication: Compression is now triggered for the
mysql.gtid_executed table when the
thread/sql/compress_gtid_table thread is launched after
the server start, and the effects are visible when the
compression process is complete. (Bug #30541799)
* Replication: Performance Schema tables could not be
accessed on a MySQL server with Group Replication that
was running under high load conditions. (Bug #30112711,
Bug #30675790)
* Replication: Internal queries from Group Replication to
the Performance Schema for statistics on local group
members failed if they occurred simultaneously with
changes to the group's membership. Locking for the
internal queries has been improved to fix the issue. (Bug
#30049349, Bug #30791583, Bug #30963553)
* Replication: In the event of an unplanned disconnection
of a replication slave from the master, the reference to
the master's dump thread might not be removed from the
list of registered slaves, in which case statements that
accessed the list of slaves would fail. The issue has now
been fixed. (Bug #29915479)
* Replication: When a partitioned table was involved, the
server did not correctly handle the situation where a row
event could not be written to the binary log due to a
lack of cache space. An appropriate error is now returned
in this situation. (Bug #29848931)
* Replication: During Group Replication's distributed
recovery process, if a joining member is unable to
complete a remote cloning operation with any donor from
the group, it uses state transfer from a donor's binary
log to retrieve all of the required data. However, if the
last attempted remote cloning operation was interrupted
and left the joining member with incomplete or no data,
an attempt at state transfer immediately afterwards could
also fail. Before attempting state transfer following a
failed remote cloning operation, Group Replication now
checks that the remote cloning operation did not reach
the stage of removing local data from the joining member.
If data was removed, the joining member leaves the group
and takes the action specified by the
group_replication_exit_state_action system variable. (Bug
#29669099, Bug #29944828)
* Replication: With the settings binlog_format=MIXED,
tx_isolation=READ-COMMITTED, and binlog_row_image=FULL,
an INSERT ... SELECT query involving a transactional
storage engine omitted any columns with a null value from
the row image written to the binary log. This happened
because when processing INSERT ... SELECT statements, the
columns were marked for inserts before the binary logging
format was selected. The issue has now been fixed. (Bug
#29110804, Bug #93423)
* Replication: Before taking certain actions, Group
Replication checks what transactions are running on the
server. Previously, the service used for this check did
not count transactions that were in the commit phase,
which could result in the action timing out. Now,
transactions that are in the commit phase are included in
the set of currently ongoing transactions. (Bug
#28327838)
* JSON: When JSON_TABLE() was used as part of an INSERT
statement in strict mode, conversion errors handled by
any ON ERROR clause could cause the INSERT to be
rejected. Since errors are handled by an ON ERROR clause,
the statement should not be rejected unless ERROR ON
ERROR is actually specified.
This issue is fixed by ignoring warnings when converting
values to the target type if NULL ON ERROR or DEFAULT ...
ON ERROR has been specified or is implied. (Bug
#30628330)
* JSON: The output from JSON_TABLE() was not always correct
when used in views. This fix corrects the following
issues:
+ Column names were not quoted, causing syntax errors
when quoting was needed for these.
+ Some column types were misreported.
+ Some column type attributes such as UNSIGNED were
lost.
+ Column character set and collation were lost.
(Bug #30263373)
* JSON: The functions JSON_SCHEMA_VALID() and
JSON_SCHEMA_VALIDATION_REPORT() formerly checked to
ensure that their arguments were convertible to JSON each
time a prepared statement including these was executed,
which was neither efficient nor necessary. Now in such
cases, the check is performed only once, when the
statement is prepared. (Bug #97878, Bug #30622327)
* Privilege requirements were checked incorrectly for
stored objects with a DEFINER that has the SYSTEM_USER
privilege. (Bug #31077699)
* A number of errors reported by Clang in the documentation
generated from the MySQL sources have been corrected.
(Bug #30956093)
* On FreeBSD, the krb5 package is a now a dependency. (Bug
#30887620)
* If a query contained multiple references to the same
common table expression (CTE) and a pseudo-comment
crossed borders of the CTE definition, the parser failed
with confusing syntax error messages. (Bug #30871301)
* For installation using Debian packages, the
/var/run/mysqld directory was not created. (Bug
#30855015, Bug #98484)
* mysqlslap did not shut down its threads properly when SQL
statements returned an error. This could result in
attempts to free already freed memory. (Bug #30850310)
* When X Plugin was attempting to add a document to a
collection as either an insertion or an update in the
case of a duplicate key, in the case where the document
failed a unique key constraint in a field other than the
primary key, the error returned by X Plugin did not state
that this was the cause of the issue. The appropriate
error is now returned. (Bug #30843865)
* An integer value generated by transformations in the
resolver was supplied to a test which expected a boolean.
(Bug #30837240)
* A query using an IN expression that accessed one or more
columns holding large string values could lead to a
memory leak. (Bug #30814171)
* Statements did not work properly when the target of a
DELETE was a common table expression. (Bug #30796015, Bug
#98330)
* Starting the server with create_admin_listener_thread
enabled and without admin_address enabled caused an
abnormal exit during the server shutdown process. (Bug
#30785609)
* When a table had both a primary key and a secondary key
on the same column, but for different lengths, the range
optimizer chose the wrong key part in the secondary index
for comparing range values. (Bug #30783011)
* In some cases, errors caused when DISTINCT was used with
an aggregate function whose argument was of an incorrect
type were not propagated correctly. (Bug #30782687)
* For replication using compression, the slave could raise
an assertion if the master was restarted. (Bug #30774692)
* For debug builds, the server could exit trying to print
an optimizer trace. (Bug #30773218, Bug #98258)
* The mysql_real_connect_nonblocking() C API function
exhibited blocking behavior. (Bug #30771233)
* With LOCK TABLES active, while processing
INFORMATION_SCHEMA queries, the server could attempt to
lock internal temporary tables (which need no locks),
causing an assertion to be raised. (Bug #30764651, Bug
#98221)
* The mysqldump internal network timeout was increased from
700 to 86400 seconds to accommodate connecting to busy or
unresponsive servers. (Bug #30755992, Bug #98203)
* Configuring with -DWITH_SASL=path/to/custom/installation
inadvertently caused libsasl to be linked into the
daemon_memcached plugin. (Bug #30755301)
* After deleting the temporary table associated with a
window function's frame buffer, the temporary table
parameter for the frame buffer was not cleaned up,
causing string buffers associated with copy fields not to
be freed properly. (Bug #30752366)
* The -libs-compat RPM package is now built with system
zlib to avoid problems with unrestricted export of
symbols in libmysqlclient.so.18. (Bug #30722389, Bug
#98130)
* The server exited histogram sampling prematurely, causing
an assertion failure. An unnecessary boolean variable
that marked the completion of a sampling operation was
removed. (Bug #30717778)
* When removing a WHERE condition because one of the
participating conditions was always false, a materialized
derived table was not cleaned up properly, resulting in a
memory leak. (Bug #30712243)
* Multiple comparisons with the same GEOMETRY value were
not always handled correctly. (Bug #30697042)
References: See also: Bug #30306306.
* MIN() and MAX() could return an incorrect value for some
queries if a WHERE clause containing an IN () subquery
was added. (Bug #30691682, Bug #98047)
* Server startup failed if MySQL Enterprise Firewall was
enabled at startup but the whitelist and user tables were
missing. (Bug #30690181)
* For prepared statements, re-execution could cause a
server exit if a cleaned-up materialized temporary table
was still being referred to. (Bug #30674598)
* The ER_WARN_DEPRECATED_SQL_CALC_FOUND_ROWS and
ER_WARN_DEPRECATED_FOUND_ROWS error messages were
incorrectly categorized in the range of messages meant to
be written to the error log. They are now correctly
categorized as messages meant to be sent to clients. The
old errors are now designated as
OBSOLETE_ER_WARN_DEPRECATED_SQL_CALC_FOUND_ROWS and
OBSOLETE_ER_WARN_DEPRECATED_FOUND_ROWS in the range of
error-log messages. (Bug #30673043)
* Some joins within subqueries where an outer query used
EXISTS or NOT EXISTS were not always handled correctly.
(Bug #30671329)
* Queries using ORDER BY constant are permitted but an
ORDER BY clause of this sort should not have any effect
on the result; such queries were not always handled
correctly. (Bug #30669493)
* A missing out-of-bounds check in wild_case_match() caused
a pointer to read out of bounds. (Bug #30668886)
* The strconvert() function was not safe for conversions
between filename and utf8_general_ci strings. (Bug
#30668847)
* Some filesorts using keys of fixed length were not always
handled correctly. (Bug #30665034)
* When performing a hash join on two string columns that
were potentially very large (in particular, BLOB columns
with PAD SPACE collations), MySQL stored the entire sort
key in the row, which impacted performance by requiring
large amounts of memory. Now only a collation-aware hash
is stored, with an added equality comparison prevent a
wrong answer, even in the event of a 64-bit hash
collision. (Bug #30664831)
* When at least two tables were joined to at least two
other tables using a semijoin, and the join optimizer
chose to use a loose scan, it was possible to place both
of the left tables below the deduplicating nested loop
iterator, leading to excessive deduplication. We fix this
by treating a loose scan across multiple tables as a
separate internal structure. (Bug #30659810)
* In unions of a const table and zero or more known-zero
expressions, derived tables of exactly one row could be
read incorrectly as having zero rows. (Bug #30655712, Bug
#97967)
* A MySQL 8.0.19 patch set an invalid INFORMATION_SCHEMA
and data dictionary version number. Assertion code was
added to prevent future version information errors. (Bug
#30645158, Bug #97948)
References: This issue is a regression of: Bug #29871530.
* When setting up the iterator tree, the optimizer now
filters away and subsequently ignores conditions which
are known to be trivially true. (Bug #30644591)
* Under some conditions, SHOW COLUMNS on a temporary MERGE
table could raise an assertion or cause a server exit.
(Bug #30640463)
References: This issue is a regression of: Bug #28811287,
Bug #92834.
* The Event Scheduler had a memory leak. (Bug #30628268)
* Using the asynchronous C API functions could result in
freeing already freed memory. (Bug #30596999, Bug #97805)
* (Bug #30594613)
* On tables containing a CHECK constraint, certain simple
queries were inefficient due to excessive memory
allocation and Performance Schema calls. (Bug #30594613)
* Under certain circumstances, a memcached command could
result in reading an uninitialized memory buffer, causing
a failure. (Bug #30592346)
* A race condition could occur between InnoDB issuing
requests for schema and table metadata while filling
INFORMATION_SCHEMA.INNODB_TABLES, and the schema being
dropped, leading to user queries on INNODB_TABLES
reporting an error. (Bug #30591967)
* Using ALTER USER to reset an account MAX_USER_CONNECTIONS
value did not take effect until all current account
connections terminated, if there were any. (Bug
#30578217, Bug #97735)
* When the optimizer sets up a weedout, it notifies all
tables that are part of the weedout that they should
provide row IDs. For confluent weedouts (weedouts
returning at most one row), the optimizer expects that
the executor handles the weedout without row IDs. In the
iterator executor, confluent weedouts are implemented
using LIMIT 1; the normal weedout iterator does not
handle confluent weedouts, and thus always expects row
IDs. In the case of a confluent weedout on the right side
of an outer join, the confluent weedout was processed as
a normal weedout, causing the iterator executor to ask
for row IDs where the tables did not supply them. Now in
such cases, the LIMIT 1 optimization is also applied.
(Bug #30566549, Bug #30282693)
* SET PERSIST could fail due to attempting to persist
variables to the wrong directory. (Bug #30561982)
* Within a stored program with an error handler defined for
the error condition of accessing a nonexistent table, the
handler was not invoked if the table was nonexistent
because it was named in a nonexistent database. (Bug
#30561920, Bug #97682)
* The duplicate weedout optimization strategy employed by
MySQL (see Optimizing IN and EXISTS Subquery Predicates
with Semijoin Transformations
(https://dev.mysql.com/doc/refman/8.0/en/semijoins.html))
uses an internal table of row IDs which it has already
seen, with a unique index on the column containing these
IDs. When the key for the unique index became too large,
which could happen with very large row IDs, the server
reverted to deduplication by hash key instead, with a
separate index (not unique) over the hash field only, as
with other temporary tables. Because the latter index was
not properly initialized, affected queries were not
executed properly and could lead to a premature exit.
(Bug #30556257)
* For debug builds, under LOCK TABLES, the server could
mishandle materialized temporary tables and raise an
assertion. (Bug #30476213, Bug #97404)
* The internal array of materialized query blocks
SELECT_LEX_UNIT::m_query_blocks_to_materialize was not
reset between executions, which meant that it pointed to
objects which were no longer valid when a prepared
statement was executed a second time, causing the second
execution to fail. (Bug #30438038)
* Altering column collations did not affect unique indexes
until a server restart. (Bug #30386119, Bug #97103)
* When using roles, the EXECUTE privilege for stored
functions was treated as a privilege for stored
procedures. As a result, it was not possible to use
EXECUTE as a role privilege for functions. (Bug
#30376231)
* A materialized subquery including a condition in which a
column value was used as input to a nondeterministic
function produced incorrect results. (Bug #30368937)
* Several fixes were applied to the InnoDB memcached
plugin. The fixes addressed potential deadlock issues,
issues related to connection list latches, and removal of
an obsolete flush mutex. (Bug #30354225)
* Strings that used the utf8mb4_0900_bin collation could
not be compared with utf8mb4 strings that used a
different collation. Now the comparison is done by using
utf8mb4_0900_bin for both strings. (Bug #30350111)
* During optimization, MySQL removes conditions in which
all arguments are considered equal; for example, 1 <> 1
is removed and replaced with false. In doing so,
conditions containing non-deterministic arguments were
also removed, which caused a condition such as RAND() <
RAND() to be considered an impossible condition. Now, the
optimizer no longer removes conditions containing
nondeterministic arguments. (Bug #30311271)
* Scheduling of events could be disturbed by removing
events. (Bug #30301356, Bug #96849)
* The Event Scheduler reported warnings for Valgrind
builds. (Bug #30301340)
* Shutting down the server while using the clone plugin
raised a Valgrind error. (Bug #30248419)
* If the mysqld-auto.cnf file was malformed, the server did
not start (expected), but did not report any error
(unexpected). (Bug #30169731, Bug #96501)
* UPDATE statements could give an inconsistent number of
rows matched (found rows) in cases where not all matched
rows were updated, depending on the reason for rows not
being updated. For example, rows not updated due to being
updated through a view with a WITH CHECK OPTION clause
were not counted as matching rows, whereas rows not
updated due to a failing CHECK CONSTRAINT were counted.
For consistency, rows that fail a WITH CHECK OPTION
clause now are counted as matching rows. (Bug #30158954)
* When restarting the MySQL server on a cloned directory,
InnoDB reported an error indicating that it could not
find a tablespace file for a statistics table that was
dropped by the server previously. (Bug #30093799)
* The server did not handle correctly a UNION in which one
of the queries contained a subquery that used ORDER BY.
(Bug #29952565)
* For INFORMATION_SCHEMA queries, a race condition could
result in multiple attempts to insert a key when updating
the dynamic statistics tables, producing a duplicate-key
error. (Bug #29948755, Bug #95929)
* SHOW CREATE VIEW could fail with an illegal mix of
collations for views defined on a function that returns a
string. (Bug #29904087)
* A query with a WHERE clause whose predicate contained a
numeric value in scientific notation was not handled
correctly.
In addition, attempting to insert a particular integer
specified as a string caused a server exit when the
string-to-integer conversion was not successful. (Bug
#29723340, Bug #30441969)
* An internal interface was added for retrieving and
parsing errors that occur on the donor MySQL server
instance (ER_CLONE_DONOR errors) and for checking if data
on the recipient has been dropped. (Bug #29682642)
* It was not possible to drop any columns from a table when
the DEFAULT value. (Bug #29661106)
* For the CONNECTION_CONTROL plugin, the Performance Schema
instrumentation used keys that were not discoverable to
the Performance Schema unless the associated code
actually executed. (Bug #29539976)
* For a nullable column c, the optimizer now recognizes
when the conditions c < c, c > c, and c <> c are always
false and need not be evaluated for every row. Thanks to
Daniel Black for the contribution. (For nonnullable
columns, the optimizer already recognized always-false
conditions.) (Bug #29115386, Bug #93642)
* Reinitialization of character sets from Index.xml could
cause a use-after-free error. (Bug #28956360, Bug #93276)
* An earlier change to reduce Performance Schema memory
instrumentation overhead had the unintended effect of
causing Group Replication performance degradation. (Bug
#28719976)
References: This issue is a regression of: Bug #27500610.
* The sys schema ps_setup_reset_to_default() procedure used
MySQL 5.7 defaults, not MySQL 8.0 defaults. (Bug
#27636611)
* Some connection encryption ciphers did not work. (Bug
#27045306)
* Previously, mysqlpump read the [mysql_dump] and [client]
groups from option files. mysqlpump now additionally
reads the [mysqlpump] group. The [mysql_dump] group is
still accepted but is deprecated. (Bug #24733245, Bug
#83144)
* For a query of the form SELECT DISTINCT ... ORDER BY ...,
when the ORDER BY was pushed down onto the first table in
the join, the result was not always sorted in the correct
order. (Bug #98217, Bug #30760534)
* The NULL indicator was not properly written for items
used as variable-length keys, such that all such items
were assumed to be not NULL, which was considered equal
to the empty string when using certain collations. One
visible effect of this issue was that ordering by an
expression using a nullable string was sometimes not
performed correctly. An example of such a query, where
column c1 contains both NULL and empty string values, is
shown here:
SELECT c1, SUBSTR(c1, 1) AS c2 FROM t ORDER BY c2;
(Bug #98035, Bug #30687020)
* A query returned inaccurate results when an expression in
a GROUP BY clause used a column name differing in case
from that used for the name of the column when the table
containing this column was created. An example of this
would be when the query used GROUP BY id although the
column name as shown in the original CREATE TABLE
statement was ID.
This occurred because, the server performed
case-sensitive comparisons of column names in expressions
with names of columns in tables. This issue is fixed by
ensuring that such comparisons are performed in a
case-insensitive fashion as expected. (Bug #97628, Bug
#98222, Bug #30541701, Bug #30761372)
* A multi-table UPDATE statement which updated a table
joined to a derived table that joined two other tables
was not optimized properly as it had been in MySQL 5.6,
instead being treated as if STRAIGHT_JOIN had been used
with the subquery creating the derived table. (Bug
#97418, Bug #30488700)
* EXPLAIN now uses hash join instead of block nested loop,
since the latter no longer exists and is replaced by a
hash join in nearly all cases. (Bug #97299, Bug
#30444550)
* The execution plan for a query that filtered on the first
column of a composite hash index wrongly used this index,
producing erroneous results. (Bug #94737, Bug #29527115)
* References to columns from tables of outer query blocks
in an ON condition of a JOIN did not work, and could be
used only in a WHERE. The fix for this problem means that
a query such as this one now works correctly:
SELECT o.order_date FROM orders o
WHERE o.order_date IN ( SELECT c.contact_name FROM customers c
INNER JOIN order_details od
ON o.order_id = od.discount );
Previously this had to be rewritten as as shown here:
SELECT o.order_date FROM orders o
WHERE o.order_date IN ( SELECT c.contact_name FROM customers c
INNER JOIN order_details od
ON 1
WHERE o.order_id = od.discount );
References to other tables of the same FROM clause as the
JOIN, as in the query SELECT * FROM t1 CROSS JOIN (t2
LEFT JOIN t3 ON t1.c=3), are not outer references and
remain forbidden. In this case, a lateral join is
required, like this: SELECT * FROM t1 JOIN LATERAL
(SELECT * FROM t2 LEFT JOIN t3 ON t1.c=3). (Bug #35242,
Bug #96946, Bug #11748138, Bug #30350696)
* There could be a mismatch between the version of OpenSSL
used to build the server and the version used for other
parts of MySQL such as libraries or plugins. This could
cause certain features not to work, such as the LDAP
authentication plugins. Now the same version of OpenSSL
is used for building everything.
* Previous work in MySQL 8.0 to optimize impossible
expressions such as a=b AND FALSE as FALSE could make for
less efficient execution when such expressions appeared
as outer join conditions, due to the fact that the join
was interpreted as a Cartesian product followed by a
filter. (Bug #8202, Bug #89739, Bug #97552, Bug
#11745046, Bug #27581277, Bug #30520749)
References: See also: Bug #98206, Bug #30756135.
Subject
Views
Written By
Posted
Sorry, you can't reply to this topic. It has been closed.
Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.