Important Change; Partitioning: User-defined partitioning of an NDBCLUSTER table without any primary key sometimes failed, and could cause mysqld to crash.

Now, if you wish to create an NDBCLUSTER table with user-defined partitioning, the table must have an explicit primary key, and all columns listed in the partitioning expression must be part of the primary key. The hidden primary key used by the NDBCLUSTER storage engine is not sufficient for this purpose. However, if the list of columns is empty (that is, the table is defined using PARTITION BY [LINEAR] KEY()), then no explicit primary key is required.

This change does not effect the partitioning of tables using any storage engine other than NDBCLUSTER. (Bug #40709)

TransactionDeadlockDetectionTimeout values less than 100 were treated as 100. This could cause scans to time out unexpectedly. (Bug #44099)

In some cases, data node restarts during a system restart could fail due to insufficient redo log space. (Bug #43156)

The file ndberror.c contained a C++-style comment, which caused builds to fail with some C compilers. (Bug #44036)

When trying to use a data node with an older version of the management server, the data node crashed on startup. (Bug #43699)

Using indexes containing variable-sized columns could lead to internal errors when the indexes were being built. (Bug #43226)

A race condition could occur when a data node failed to restart just before being included in the next global checkpoint. This could cause other data nodes to fail. (Bug #43888)

When a data node had written its GCI marker to the first page of a megabyte, and that node was later killed during restart after having processed that page (marker) but before completing a LCP, the data node could fail with file system errors. (Bug #44952)

References: See also Bug #42564, Bug #44291.

Inspection of the code revealed that several assignment operators (=) were used in place of comparison operators (==) in DbdihMain.cpp. (Bug #44567)

References: See also Bug #44570.

An internal NDB API buffer was not properly initialized. (Bug #44977)

Some queries using combinations of logical and comparison operators on an indexed column in the WHERE clause could fail with the error Got error 4541 'IndexBound has no bound information' from NDBCLUSTER. (Bug #42857)

ndb_restore --print_data did not handle DECIMAL columns correctly. (Bug #37171)

It was possible for NDB API applications to insert corrupt data into the database, which could subquently lead to data node crashes. Now, stricter checking is enforced on input data for inserts and updates. (Bug #44132)

When aborting an operation involving both an insert and a delete, the insert and delete were aborted separately. This was because the transaction coordinator did not know that the operations affected on same row, and, in the case of a committed-read (tuple or index) scan, the abort of the insert was performed first, then the row was examined after the insert was aborted but before the delete was aborted. In some cases, this would leave the row in a inconsistent state. This could occur when a local checkpoint was performed during a backup. This issue did not affect primary ley operations or scans that used locks (these are serialized).

After this fix, for ordered indexes, all operations that follow the operation to be aborted are now also aborted.

It was theoretically possible for the value of a nonexistent column to be read as NULL, rather than causing an error. (Bug #27843)

The output of ndbd --help did not provide clear information about the program's --initial and --initial-start options. (Bug #28905)

Partitioning; Disk Data: An NDB table created with a very large value for the MAX_ROWS option could—if this table was dropped and a new table with fewer partitions, but having the same table ID, was created—cause ndbd to crash when performing a system restart. This was because the server attempted to examine each partition whether or not it actually existed. (Bug #45154)

References: See also Bug #58638.

Disk Data: During a checkpoint, restore points are created for both the on-disk and in-memory parts of a Disk Data table. Under certain rare conditions, the in-memory restore point could include or exclude a row that should have been in the snapshot. This would later lead to a crash during or following recovery. (Bug #41915)

References: See also Bug #47832.

Disk Data: When a log file group had an undo log file whose size was too small, restarting data nodes failed with Read underflow errors.

As a result of this fix, the minimum permitted INTIAL_SIZE for an undo log file is now 1M (1 megabyte). (Bug #29574)

Disk Data: This fix supersedes and improves on an earlier fix made for this bug in MySQL 5.1.18. (Bug #24521)

Cluster API: The default NdbRecord structures created by NdbDictionary could have overlapping null bits and data fields. (Bug #43590)

Cluster API: Ordered index scans using NdbRecord formerly expressed a BoundEQ range as separate lower and upper bounds, resulting in 2 copies of the column values being sent to the NDB kernel.

Now, when a range is specified by NdbScanOperation::setBound(), the passed pointers, key lengths, and inclusive bits are compared, and only one copy of the equal key columns is sent to the kernel. This makes such operations more efficient, as half the amount of KeyInfo is now sent for a BoundEQ range as before. (Bug #38793)

Cluster API: When performing insert or write operations, NdbRecord permits key columns to be specified in both the key record and in the attribute record. Only one key column value for each key column should be sent to the NDB kernel, but this was not guaranteed. This is now ensured as follows: For insert and write operations, key column values are taken from the key record; for scan takeover update operations, key column values are taken from the attribute record. (Bug #42238)

Cluster API: If the largest offset of a RecordSpecification used for an NdbRecord object was for the NULL bits (and thus not a column), this offset was not taken into account when calculating the size used for the RecordSpecification. This meant that the space for the NULL bits could be overwritten by key or other information. (Bug #43891)

Important Change: Formerly, when the management server failed to create a transporter for a data node connection, net_write_timeout seconds elapsed before the data node was actually permitted to disconnect. Now in such cases the disconnection occurs immediately. (Bug #41965)

References: See also Bug #41713.

Disk Data: It is now possible to specify default locations for Disk Data data files and undo log files, either together or separately, using the data node configuration parameters FileSystemPathDD, FileSystemPathDataFiles, and FileSystemPathUndoFiles. For information about these configuration parameters, see Disk Data file system parameters.

It is also now possible to specify a log file group, tablespace, or both, that is created when the cluster is started, using the InitialLogFileGroup and InitialTablespace data node configuration parameters. For information about these configuration parameters, see Disk Data object creation parameters.

Performance: Updates of the SYSTAB_0 system table to obtain a unique identifier did not use transaction hints for tables having no primary key. In such cases the NDB kernel used a cache size of 1. This meant that each insert into a table not having a primary key required an update of the corresponding SYSTAB_0 entry, creating a potential performance bottleneck.

With this fix, inserts on NDB tables without primary keys can be under some conditions be performed up to 100% faster than previously. (Bug #39268)

Packaging: Packages for MySQL Cluster were missing the libndbclient.so and libndbclient.a files. (Bug #42278)

Partitioning: Executing ALTER TABLE ... REORGANIZE PARTITION on an NDBCLUSTER table having only one partition caused mysqld to crash. (Bug #41945)

References: See also Bug #40389.

Cluster API: Failed operations on BLOB and TEXT columns were not always reported correctly to the originating SQL node. Such errors were sometimes reported as being due to timeouts, when the actual problem was a transporter overload due to insufficient buffer space. (Bug #39867, Bug #39879)

It was theoretically possible for stale data to be read from NDBCLUSTER tables when the transaction isolation level was set to ReadCommitted. (Bug #40543)

Issuing SHOW GLOBAL STATUS LIKE 'NDB%' before mysqld had connected to the cluster caused a segmentation fault. (Bug #42458)

In the event that a MySQL Cluster backup failed due to file permissions issues, conflicting reports were issued in the management client. (Bug #34526)

When a cluster backup failed with Error 1304 (Node node_id1: Backup request from node_id2 failed to start), no clear reason for the failure was provided.

As part of this fix, MySQL Cluster now retries backups in the event of sequence errors. (Bug #42354)

References: See also Bug #22698.

Events logged after setting ALL CLUSTERLOG STATISTICS=15 in the management client did not always include the node ID of the reporting node. (Bug #39839)

Issuing SHOW ENGINE NDBCLUSTER STATUS on an SQL node before the management server had connected to the cluster caused mysqld to crash. (Bug #42264)

If a transaction was aborted during the handling of a data node failure, this could lead to the later handling of an API node failure not being completed. (Bug #41214)

Start phase reporting was inconsistent between the management client and the cluster log. (Bug #39667)

If the master node failed during a global checkpoint, it was possible in some circumstances for the new master to use an incorrect value for the global checkpoint index. This could occur only when the cluster used more than one node group. (Bug #41469)

Transaction failures took longer to handle than was necessary.

When a data node acting as transaction coordinator (TC) failed, the surviving data nodes did not inform the API node initiating the transaction of this until the failure had been processed by all protocols. However, the API node needed only to know about failure handling by the transaction protocol—that is, it needed to be informed only about the TC takeover process. Now, API nodes (including MySQL servers acting as cluster SQL nodes) are informed as soon as the TC takeover is complete, so that it can carry on operating more quickly. (Bug #40697)

Restoring a MySQL Cluster from a dump made using mysqldump failed due to a spurious error: Can't execute the given command because you have active locked tables or an active transaction. (Bug #40346)

Memory leaks could occur in handling of strings used for storing cluster metadata and providing output to users. (Bug #38662)

Data node failures that occurred before all data nodes had connected to the cluster were not handled correctly, leading to additional data node failures. (Bug #42422)

A data node failure that occurred between calls to NdbIndexScanOperation::readTuples(SF_OrderBy) and NdbTransaction::execute() was not correctly handled; a subsequent call to nextResult() caused a null pointer to be deferenced, leading to a segfault in mysqld. (Bug #42545)

Triggers on NDBCLUSTER tables caused such tables to become locked. (Bug #42751)

References: See also Bug #16229, Bug #18135.

A maximum of 11 TUP scans were permitted in parallel. (Bug #42084)

Backup IDs greater than 2³¹ were not handled correctly, causing negative values to be used in backup directory names and printouts. (Bug #43042)

When using multiple management servers and starting several API nodes (possibly including one or more SQL nodes) whose connectstrings listed the management servers in different order, it was possible for 2 API nodes to be assigned the same node ID. When this happened it was possible for an API node not to get fully connected, consequently producing a number of errors whose cause was not easily recognizable. (Bug #42973)

In some cases, NDB did not check correctly whether tables had changed before trying to use the query cache. This could result in a crash of the debug MySQL server. (Bug #40464)

Issuing EXIT in the management client sometimes caused the client to hang. (Bug #40922)

ndb_error_reporter worked correctly only with GNU tar. (With other versions of tar, it produced empty archives.) (Bug #42753)

When performing more than 32 index or tuple scans on a single fragment, the scans could be left hanging. This caused unnecessary timeouts, and in addition could possibly lead to a hang of an LCP. (Bug #42559)

References: This bug is a regression of Bug #42084.

Given a MySQL Cluster containing no data (that is, whose data nodes had all been started using --initial, and into which no data had yet been imported) and having an empty backup directory, executing START BACKUP with a user-specified backup ID caused the data nodes to crash. (Bug #41031)

Trying to execute an ALTER ONLINE TABLE ... ADD COLUMN statement while inserting rows into the table caused mysqld to crash. (Bug #41905)

When using ndbmtd, NDB kernel threads could hang while trying to start the data nodes with LockPagesInMainMemory set to 1. (Bug #43021)

During node failure handling (of a data node other than the master), there was a chance that the master was waiting for a GCP_NODEFINISHED signal from the failed node after having received it from all other data nodes. If this occurred while the failed node had a transaction that was still being committed in the current epoch, the master node could crash in the DBTC kernel block when discovering that a transaction actually belonged to an epoch which was already completed. (Bug #41295)

The MySQL Query Cache did not function correctly with NDBCLUSTER tables containing TEXT columns. (Bug #39295)

An abort path in the DBLQH kernel block failed to release a commit acknowledgment marker. This meant that, during node failure handling, the local query handler could be added multiple times to the marker record which could lead to additional node failures due an array overflow. (Bug #41296)

A segfault in Logger::Log caused ndbd to hang indefinitely. This fix improves on an earlier one for this issue, first made in MySQL Cluster NDB 6.2.16 and MySQL Cluster NDB 6.3.17. (Bug #39180)

References: See also Bug #38609.

O_DIRECT was incorrectly disabled when making MySQL Cluster backups. (Bug #40205)

Redo log creation was very slow on some platforms, causing MySQL Cluster to start more slowly than necessary with some combinations of hardware and operating system. This was due to all write operations being synchronized to disk while creating a redo log file. Now this synchronization occurs only after the redo log has been created. (Bug #40734)

API nodes disconnected too agressively from cluster when data nodes were being restarted. This could sometimes lead to the API node being unable to access the cluster at all during a rolling restart. (Bug #41462)

A duplicate key or other error raised when inserting into an NDBCLUSTER table caused the current transaction to abort, after which any SQL statement other than a ROLLBACK failed. With this fix, the NDBCLUSTER storage engine now performs an implicit rollback when a transaction is aborted in this way; it is no longer necessary to issue an explicit ROLLBACK statement, and the next statement that is issued automatically begins a new transaction.

(Bug #32656)

References: See also Bug #47654.

Error messages for NDBCLUSTER error codes 1224 and 1227 were missing. (Bug #28496)

Disk Data: Issuing concurrent CREATE TABLESPACE, ALTER TABLESPACE, CREATE LOGFILE GROUP, or ALTER LOGFILE GROUP statements on separate SQL nodes caused a resource leak that led to data node crashes when these statements were used again later. (Bug #40921)

Disk Data: Creation of a tablespace data file whose size was greater than 4 GB failed silently on 32-bit platforms. (Bug #37116)

References: See also Bug #29186.

Disk Data: O_SYNC was incorrectly disabled on platforms that do not support O_DIRECT. This issue was noted on Solaris but could have affected other platforms not having O_DIRECT capability. (Bug #34638)

Disk Data: Creating a Disk Data tablespace with a very large extent size caused the data nodes to fail. The issue was observed when using extent sizes of 100 MB and larger. (Bug #39096)

Disk Data: It was not possible to add an in-memory column online to a table that used a table-level or column-level STORAGE DISK option. The same issue prevented ALTER ONLINE TABLE ... REORGANIZE PARTITION from working on Disk Data tables. (Bug #42549)

Disk Data: Disk-based variable-length columns were not always handled like their memory-based equivalents, which could potentially lead to a crash of cluster data nodes. (Bug #39645)

Disk Data: Trying to execute a CREATE LOGFILE GROUP statement using a value greater than 150M for UNDO_BUFFER_SIZE caused data nodes to crash.

As a result of this fix, the upper limit for UNDO_BUFFER_SIZE is now 600M; attempting to set a higher value now fails gracefully with an error. (Bug #34102)

References: See also Bug #36702.

Disk Data: When attempting to create a tablespace that already existed, the error message returned was Table or index with given name already exists. (Bug #32662)

Disk Data: Starting a cluster under load such that Disk Data tables used most of the undo buffer could cause data node failures.

The fix for this bug also corrected an issue in the LGMAN kernel block where the amount of free space left in the undo buffer was miscalculated, causing buffer overruns. This could cause records in the buffer to be overwritten, leading to problems when restarting data nodes. (Bug #28077)

Disk Data: Using a path or file name longer than 128 characters for Disk Data undo log files and tablespace data files caused a number of issues, including failures of CREATE LOGFILE GROUP, ALTER LOGFILE GROUP, CREATE TABLESPACE, and ALTER TABLESPACE statements, as well as crashes of management nodes and data nodes.

With this fix, the maximum length for path and file names used for Disk Data undo log files and tablespace data files is now the same as the maximum for the operating system. (Bug #31769, Bug #31770, Bug #31772)

Disk Data: Attempting to perform a system restart of the cluster where there existed a logfile group without and undo log files caused the data nodes to crash.

(Bug #17614)

Cluster API: The NDB API example programs included in MySQL Cluster source distributions failed to compile. (Bug #37491)

References: See also Bug #40238.

Cluster API: It was not always possible to access parent objects directly from NdbBlob, NdbOperation, and NdbScanOperation objects. To alleviate this problem, a new getNdbOperation() method has been added to NdbBlob and new getNdbTransaction() methods have been added to NdbOperation and NdbScanOperation. In addition, a const variant of NdbOperation::getErrorLine() is now also available. (Bug #40242)

Cluster API: NdbScanOperation::getBlobHandle() failed when used with incorrect column names or numbers. (Bug #40241)

Cluster API: Some error messages from ndb_mgmd contained newline (\n) characters. This could break the MGM API protocol, which uses the newline as a line separator. (Bug #43104)

Cluster API: The MGM API reset error codes on management server handles before checking them. This meant that calling an MGM API function with a null handle caused applications to crash. (Bug #40455)

Cluster API: When using an ordered index scan without putting all key columns in the read mask, this invalid use of the NDB API went undetected, which resulted in the use of uninitialized memory. (Bug #42591)

Cluster API: mgmapi.h contained constructs which only worked in C++, but not in C. (Bug #27004)

It is no longer a requirement for database autodiscovery that an SQL node already be connected to the cluster at the time that a database is created on another SQL node. It is no longer necessary to issue CREATE DATABASE (or CREATE SCHEMA) statements on an SQL node joining the cluster after a database is created for the new SQL node to see the database and any NDBCLUSTER tables that it contains. (Bug #39612)

Renaming an NDBCLUSTER table on one SQL node, caused a trigger on this table to be deleted on another SQL node. (Bug #36658)

Failure to parse configuration parameters could cause a memory leak in the NDB log parser. (Bug #38380)

Attempting to add a UNIQUE INDEX twice to an NDBCLUSTER table, then deleting rows from the table could cause the MySQL Server to crash. (Bug #35599)

Setting a low value of MaxNoOfLocalScans (< 100) and performing a large number of (certain) scans could cause the Transaction Coordinator to run out of scan fragment records, and then crash. Now when this resource is exhausted, the cluster returns Error 291 (Out of scanfrag records in TC (increase MaxNoOfLocalScans)) instead. (Bug #39549)

A failed connection to the management server could cause a resource leak in ndb_mgmd. (Bug #38424)

ndb_restore failed with a floating point exception due to a division by zero error when trying to restore certain data files. (Bug #38520)

ndb_restore failed when a single table was specified. (Bug #33801)

ndb_mgmd sometimes performed unnecessary network I/O with the client. This in combination with other factors led to long-running threads that were attempting to write to clients that no longer existed. (Bug #38563)

When restarting a data node, an excessively long shutdown message could cause the node process to crash. (Bug #38580)

Creating a unique index on an NDBCLUSTER table caused a memory leak in the NDB subscription manager (SUMA) which could lead to mysqld hanging, due to the fact that the resource shortage was not reported back to the NDB kernel correctly. (Bug #39518)

References: See also Bug #39450.

Starting the MySQL Server with the --ndbcluster option plus an invalid command-line option (for example, using mysqld --ndbcluster --foobar) caused it to hang while shutting down the binlog thread. (Bug #39635)

An invalid memory access caused the management server to crash on Solaris Sparc platforms. (Bug #38628)

ndb_mgmd failed to start on older Linux distributions (2.4 kernels) that did not support e-polling. (Bug #38592)

A segfault in Logger::Log caused ndbd to hang indefinitely. (Bug #38609)

An invalid path specification caused mysql-test-run.pl to fail. (Bug #39026)

After a forced shutdown and initial restart of the cluster, it was possible for SQL nodes to retain .frm files corresponding to NDBCLUSTER tables that had been dropped, and thus to be unaware that these tables no longer existed. In such cases, attempting to re-create the tables using CREATE TABLE IF NOT EXISTS could fail with a spurious Table ... doesn't exist error. (Bug #37921)

Unique identifiers in tables having no primary key were not cached. This fix has been observed to increase the efficiency of INSERT operations on such tables by as much as 50%. (Bug #39267)

Dropping and then re-creating a database on one SQL node caused other SQL nodes to hang. (Bug #39613)

MgmtSrvr::allocNodeId() left a mutex locked following an Ambiguity for node if %d error. (Bug #39158)

Heavy DDL usage caused the mysqld processes to hang due to a timeout error (NDB error code 266). (Bug #39885)

During transactional coordinator takeover (directly after node failure), the LQH finding an operation in the LOG_COMMIT state sent an LQH_TRANS_CONF signal twice, causing the TC to fail. (Bug #38930)

Executing EXPLAIN SELECT on an NDBCLUSTER table could cause mysqld to crash. (Bug #39872)

GCP_COMMIT did not wait for transaction takeover during node failure. This could cause GCP_SAVE_REQ to be executed too early. This could also cause (very rarely) replication to skip rows. (Bug #30780)

Cluster API: The NdbScanOperation::readTuples() method could be called multiple times without error. (Bug #38717)

Cluster API: Creating an NdbScanFilter object using an NdbScanOperation object that had not yet had its readTuples() method called resulted in a crash when later attempting to use the NdbScanFilter. (Bug #37986)

Cluster API: Certain Multi-Range Read scans involving IS NULL and IS NOT NULL comparisons failed with an error in the NDB local query handler. (Bug #38204)

Cluster API: Passing a value greater than 65535 to NdbInterpretedCode::add_val() and NdbInterpretedCode::sub_val() caused these methods to have no effect. (Bug #39536)

Cluster API: Executing an NdbRecord interpreted delete created with an ANYVALUE option caused the transaction to abort. (Bug #37672)

Cluster API: Accessing the debug version of libndbclient using dlopen() resulted in a segmentation fault. (Bug #35927)

Cluster API: Problems with the public headers prevented NDB applications from being built with warnings turned on. (Bug #38177)

Added the MaxBufferedEpochs data node configuration parameter, which controls the maximum number of unprocessed epochs by which a subscribing node can lag. Subscribers which exceed this number are disconnected and forced to reconnect.

See Defining MySQL Cluster Data Nodes, for more information.

Incompatible Change: The UPDATE statement permitted NULL to be assigned to NOT NULL columns (the implicit default value for the column data type was assigned). This was changed so that on error occurs.

This change was reverted, because the original report was determined not to be a bug: Assigning NULL to a NOT NULL column in an UPDATE statement should produce an error only in strict SQL mode and set the column to the implicit default with a warning otherwise, which was the original behavior. See Data Type Default Values, and Bug #39265. (Bug #33699)

Incompatible Change: Previously, the parser accepted the ODBC { OJ ... LEFT OUTER JOIN ...} syntax for writing left outer joins. The parser now permits { OJ ... } to be used to write other types of joins, such as INNER JOIN or RIGHT OUTER JOIN. This helps with compatibility with some third-party applications, but is not official ODBC syntax.

A consequence of this change is that the parser no longer permits nested { OJ ... } constructs (which are not legal ODBC syntax, anyway). Queries that use such constructs should be rewritten. For example, this query is now produces an error:

SELECT * FROM
   {OJ
      {OJ a LEFT OUTER JOIN b ON a.a1=b.a1}
      LEFT OUTER JOIN c ON b.b1 = c.b1};

That can be replaced by any of the following rewrites:

SELECT * FROM
    {OJ a LEFT OUTER JOIN b
          LEFT OUTER JOIN c ON b.b1 = c.b1 ON a.a1=b.a1};

SELECT * FROM
   {OJ a LEFT OUTER JOIN b ON a.a1=b.a1
         LEFT OUTER JOIN c ON b.b1 = c.b1};

SELECT * FROM
     a LEFT OUTER JOIN b ON a.a1=b.a1 LEFT OUTER JOIN c ON b.b1 = c.b1;

The first two are legal according to ODBC, and you nest the joins inside a single { OJ ...} clause. The third is standard SQL syntax, without ODBC decoration. It can be used with parentheses to emphasize the evaluation order:

SELECT * FROM
     ((a LEFT OUTER JOIN b ON a.a1=b.a1)
         LEFT OUTER JOIN c ON b.b1 = c.b1);

(Bug #28317)

Cluster API: Using NDB API applications from older MySQL Cluster versions with libndbclient from newer ones caused the cluster to fail. (Bug #36124)

Cluster API: Scans having no bounds set were handled incorrectly. (Bug #35876)

Cluster API: Closing a scan before it was executed caused the application to segfault. (Bug #36375)

A node failure during an initial node restart followed by another node start could cause the master data node to fail, because it incorrectly gave the node permission to start even if the invalidated node's LCP was still running. (Bug #34702)

The failure of a DDL statement could sometimes lead to node failures when attempting to execute subsequent DDL statements. (Bug #34160)

In certain rare circumstances, a race condition could occur between an aborted insert and a delete leading a data node crash. (Bug #34260)

Statements executing multiple inserts performed poorly on NDB tables having AUTO_INCREMENT columns. (Bug #33534)

Performing many SQL statements on NDB tables while in autocommit mode caused a memory leak in mysqld. (Bug #34275)

Upgrades of a cluster using while a DataMemory setting in excess of 16 GB caused data nodes to fail. (Bug #34378)

Extremely long SELECT statements (where the text of the statement was in excess of 50000 characters) against NDB tables returned empty results. (Bug #34107)

When configured with NDB support, MySQL failed to compile using gcc 4.3 on 64bit FreeBSD systems. (Bug #34169)

Multi-table updates using ordered indexes during handling of node failures could cause other data nodes to fail. (Bug #34216)

Having tables with a great many columns could cause Cluster backups to fail. (Bug #30172)

Transaction atomicity was sometimes not preserved between reads and inserts under high loads. (Bug #31477)

The ndb_waiter utility polled ndb_mgmd excessively when obtaining the status of cluster data nodes. (Bug #32025)

References: See also Bug #32023.

Disk Data; Cluster Replication: Statements violating unique keys on Disk Data tables (such as attempting to insert NULL into a NOT NULL column) could cause data nodes to fail. When the statement was executed from the binlog, this could also result in failure of the slave cluster. (Bug #34118)

Disk Data: Updating in-memory columns of one or more rows of Disk Data table, followed by deletion of these rows and re-insertion of them, caused data node failures. (Bug #33619)

Important Change; Cluster API: Because NDB_LE_MemoryUsage.page_size_kb shows memory page sizes in bytes rather than kilobytes, it has been renamed to page_size_bytes. The name page_size_kb is now deprecated and thus subject to removal in a future release, although it currently remains supported for reasons of backward compatibility. See The Ndb_logevent_type Type, for more information about NDB_LE_MemoryUsage. (Bug #30271)

A periodic failure to flush the send buffer by the NDB TCP transporter could cause a unnecessary delay of 10 ms between operations. (Bug #34005)

High numbers of insert operations, delete operations, or both could cause NDB error 899 (Rowid already allocated) to occur unnecessarily. (Bug #34033)

Some tuple scans caused the wrong memory page to be accessed, leading to invalid results. This issue could affect both in-memory and Disk Data tables. (Bug #33739)

A race condition could occur (very rarely) when the release of a GCI was followed by a data node failure. (Bug #33793)

Issuing an INSERT ... ON DUPLICATE KEY UPDATE concurrently with or following a TRUNCATE TABLE statement on an NDB table failed with NDB error 4350 Transaction already aborted. (Bug #29851)

The server failed to reject properly the creation of an NDB table having an unindexed AUTO_INCREMENT column. (Bug #30417)

It was possible in config.ini to define cluster nodes having node IDs greater than the maximum permitted value. (Bug #28298)

The Cluster backup process could not detect when there was no more disk space and instead continued to run until killed manually. Now the backup fails with an appropriate error when disk space is exhausted. (Bug #28647)

Cluster API: When reading a BIT(64) value using NdbOperation:getValue(), 12 bytes were written to the buffer rather than the expected 8 bytes. (Bug #33750)

Cluster API: Transactions containing inserts or reads would hang during NdbTransaction::execute() calls made from NDB API applications built against a MySQL Cluster version that did not support micro-GCPs accessing a later version that supported micro-GCPs. This issue was observed while upgrading from MySQL Cluster NDB 6.1.23 to MySQL Cluster NDB 6.2.10 when the API application built against the earlier version attempted to access a data node already running the later version, even after disabling micro-GCPs by setting TimeBetweenEpochs equal to 0. (Bug #33895)

Partitioning: When partition pruning on an NDB table resulted in an ordered index scan spanning only one partition, any descending flag for the scan was wrongly discarded, causing ORDER BY DESC to be treated as ORDER BY ASC, MAX() to be handled incorrectly, and similar problems. (Bug #33061)

When all data and SQL nodes in the cluster were shut down abnormally (that is, other than by using STOP in the cluster management client), ndb_mgm used excessive amounts of CPU. (Bug #33237)

When using micro-GCPs, if a node failed while preparing for a global checkpoint, the master node would use the wrong GCI. (Bug #32922)

Under some conditions, performing an ALTER TABLE on an NDBCLUSTER table failed with a Table is full error, even when only 25% of DataMemory was in use and the result should have been a table using less memory (for example, changing a VARCHAR(100) column to VARCHAR(80)). (Bug #32670)

Added the ndb_mgm client command DUMP 8011, which dumps all subscribers to the cluster log. See DUMP 8011, for more information.

Adding a new TINYTEXT column to an NDB table which used COLUMN_FORMAT = DYNAMIC, and when binary logging was enabled, caused all cluster mysqld processes to crash. (Bug #30213)

The failure of a master node could lead to subsequent failures in local checkpointing. (Bug #32160)

A restart of the cluster failed when more than 1 REDO phase was in use. (Bug #22696)

High numbers of API nodes on a slow or congested network could cause connection negotiation to time out prematurely, leading to the following issues:

(Bug #32359)

After adding a new column of one of the TEXT or BLOB types to an NDB table which used COLUMN_FORMAT = DYNAMIC, it was no longer possible to access or drop the table using SQL. (Bug #30205)

A local checkpoint could sometimes be started before the previous LCP was restorable from a global checkpoint. (Bug #32519)

The output of the ndb_mgm client SHOW and STATUS commands now indicates when the cluster is in single user mode. (Bug #27999)

The NDB management client command node_id REPORT MEMORY provided no output when node_id was the node ID of a management or API node. Now, when this occurs, the management client responds with Node node_id: is not a data node. (Bug #29485)

The NDB storage engine code was not safe for strict-alias optimization in gcc 4.2.1. (Bug #31761)

An insert or update with combined range and equality constraints failed when run against an NDB table with the error Got unknown error from NDB. An example of such a statement would be UPDATE t1 SET b = 5 WHERE a IN (7,8) OR a >= 10;. (Bug #31874)

An error with an if statement in sql/ha_ndbcluster.cc could potentially lead to an infinite loop in case of failure when working with AUTO_INCREMENT columns in NDB tables. (Bug #31810)

UPDATE IGNORE could sometimes fail on NDB tables due to the use of unitialized data when checking for duplicate keys to be ignored. (Bug #25817)

Following an upgrade, ndb_mgmd would fail with an ArbitrationError. (Bug #31690)

In a cluster running in diskless mode and with arbitration disabled, the failure of a data node during an insert operation caused other data node to fail. (Bug #31980)

Performing DELETE operations after a data node had been shut down could lead to inconsistent data following a restart of the node. (Bug #26450)

A cluster restart could sometimes fail due to an issue with table IDs. (Bug #30975)

A node failure during a local checkpoint could lead to a subsequent failure of the cluster during a system restart. (Bug #31257)

It was possible in some cases for a node group to be “lost” due to missed local checkpoints following a system restart. (Bug #31525)

Transaction timeouts were not handled well in some circumstances, leading to excessive number of transactions being aborted unnecessarily. (Bug #30379)

The cluster log was formatted inconsistently and contained extraneous newline characters. (Bug #25064)

NDB tables having names containing nonalphanumeric characters (such as “$”) were not discovered correctly. (Bug #31470)

ndb_mgm --help did not display any information about the -a option. (Bug #29509)

In some cases, the cluster managment server logged entries multiple times following a restart of ndb_mgmd. (Bug #29565)

Mapping of NDB error codes to MySQL storage engine error codes has been improved. (Bug #28423)

Partitioning: EXPLAIN PARTITIONS reported partition usage by queries on NDB tables according to the standard MySQL hash function than the hash function used in the NDB storage engine. (Bug #29550)

Attempting to restore a backup made on a cluster host using one endian to a machine using the other endian could cause the cluster to fail. (Bug #29674)

An insufficiently descriptive and potentially misleading Error 4006 (Connect failure - out of connection objects...) was produced when either of the following two conditions occurred:

The description of the --print option provided in the output from ndb_restore --help was incorrect. (Bug #27683)

When an NDB event was left behind but the corresponding table was later recreated and received a new table ID, the event could not be dropped. (Bug #30877)

Restoring a backup made on a cluster host using one endian to a machine using the other endian failed for BLOB and DATETIME columns. (Bug #27543, Bug #30024)

Online ADD COLUMN, ADD INDEX, and DROP INDEX operations can now be performed explicitly for NDB tables—that is, without copying or locking of the affected tables—using ALTER ONLINE TABLE. Indexes can also be created and dropped online using CREATE INDEX and DROP INDEX, respectively, using the ONLINE keyword.

You can force operations that would otherwise be performed online to be done offline using the OFFLINE keyword.

Renaming of tables and columns for NDB and MyISAM tables is performed in place without table copying.

For more information, see ALTER TABLE Online Operations in MySQL Cluster, CREATE INDEX Syntax, and DROP INDEX Syntax.

It is now possible to control whether fixed-width or variable-width storage is used for a given column of an NDB table by means of the COLUMN_FORMAT specifier as part of the column's definition in a CREATE TABLE or ALTER TABLE statement.

It is also possible to control whether a given column of an NDB table is stored in memory or on disk, using the STORAGE specifier as part of the column's definition in a CREATE TABLE or ALTER TABLE statement.

For permitted values and other information about COLUMN_FORMAT and STORAGE, see CREATE TABLE Syntax.

A new cluster management server startup option --bind-address makes it possible to restrict management client connections to ndb_mgmd to a single host and port. For more information, see ndb_mgmd — The MySQL Cluster Management Server Daemon.

The following improvements have been made in the ndb_size.pl utility:

Support for a --socket option was also added.

For more information, see ndb_size.pl — NDBCLUSTER Size Requirement Estimator. (Bug #28683, Bug #28253)

Using the --ndb-cluster-connection-pool option for mysqld caused DDL statements to be executed twice. (Bug #30598)

A file system close operation could fail during a node or system restart. (Bug #30646)

When handling BLOB columns, the addition of read locks to the lock queue was not handled correctly. (Bug #30764)

ndb_size.pl failed on tables with FLOAT columns whose definitions included commas (for example, FLOAT(6,2)). (Bug #29228)

A query using joins between several large tables and requiring unique index lookups failed to complete, eventually returning Unknown Error after a very long period of time. This occurred due to inadequate handling of instances where the Transaction Coordinator ran out of TransactionBufferMemory, when the cluster should have returned NDB error code 4012 (Request ndbd time-out). (Bug #28804)

Reads on BLOB columns were not locked when they needed to be to guarantee consistency. (Bug #29102)

References: See also Bug #31482.

An attempt to perform a SELECT ... FROM INFORMATION_SCHEMA.TABLES whose result included information about NDB tables for which the user had no privileges crashed the MySQL Server on which the query was performed. (Bug #26793)

When creating an NDB table with a column that has COLUMN_FORMAT = DYNAMIC, but the table itself uses ROW_FORMAT=FIXED, the table is considered dynamic, but any columns for which the row format is unspecified default to FIXED. Now in such cases the server issues the warning Row format FIXED incompatible with dynamic attribute column_name. (Bug #30276)

Discovery of NDB tables did not work correctly with INFORMATION_SCHEMA. (Bug #30667)

Cluster API: A call to CHECK_TIMEDOUT_RET() in mgmapi.cpp should have been a call to DBUG_CHECK_TIMEDOUT_RET(). (Bug #30681)

When restarting a data node, queries could hang during that node's start phase 5, and continue only after the node had entered phase 6. (Bug #29364)

Replica redo logs were inconsistently handled during a system restart. (Bug #29354)

Disk Data: The number of free extents was incorrectly reported for some tablespaces. (Bug #28642)

Disk Data: Performing Disk Data schema operations during a node restart could cause forced shutdowns of other data nodes. (Bug #29501)

Disk Data: Disk data meta-information that existed in ndbd might not be visible to mysqld. (Bug #28720)

Important Change: The TimeBetweenWatchdogCheckInitial configuration parameter was added to enable setting of a separate watchdog timeout for memory allocation during startup of the data nodes. (Bug #28899)

Important Change; Cluster API: A new NdbRecord object has been added to the NDB API. This object provides mapping to a record stored in NDB.

A new data node configuration parameter BackupReportFrequency now makes it possible to cause the management client to provide status reports at regular intervals as well as for such reports to be written to the cluster log (depending on cluster event logging levels).

A new memory allocator has been implemented for the NDB kernel, which allocates memory to tables 32K page by 32K page rather than allocating it in variable-sized chunks as previously. This removes much of the memory overhead that was associated with the old memory allocator.

auto_increment_increment and auto_increment_offset are now supported for NDB tables. (Bug #26342)

ndb_restore now provides running reports of its progress when restoring a backup. In addition, a complete report status report on the backup is written to the cluster log.

Reporting functionality has been significantly enhanced in this release:

A REPORT BackupStatus command has been added in the cluster management client. This command enables you to obtain a backup status report at any time during a backup. For more about this command, see Commands in the MySQL Cluster Management Client.

A new configuration parameter ODirect causes NDB to attempt using O_DIRECT writes for LCP, backups, and redo logs, often lowering CPU usage.

It is now possible to set the maximum size of the allocation unit for table memory using the MaxAllocate configuration parameter. (Bug #29044)

Attempting to restore a NULL row to a VARBINARY column caused ndb_restore to fail. (Bug #29103)

When a node failed to respond to a COPY_GCI signal as part of a global checkpoint, the master node was killed instead of the node that actually failed. (Bug #29331)

In the event that two data nodes in the same node group and participating in a GCP crashed before they had written their respective P0.sysfile files, QMGR could refuse to start, issuing an invalid Insufficient nodes for restart error instead. (Bug #29167)

LCP files were not removed following an initial system restart. (Bug #28726)

When shutting down mysqld, the NDB binlog process was not shut down before log cleanup began. (Bug #28949)

ndb_error_reporter now preserves timestamps on files. (Bug #29074)

Having large amounts of memory locked caused swapping to disk. (Bug #28751)

The wrong data pages were sometimes invalidated following a global checkpoint. (Bug #29067)

A corrupt schema file could cause a File already open error. (Bug #28770)

A fast global checkpoint under high load with high usage of the redo buffer caused data nodes to fail. (Bug #28653)

If at least 2 files were involved in REDO invalidation, then file 0 of page 0 was not updated and so pointed to an invalid part of the redo log. (Bug #29057)

A query having a large IN(...) or NOT IN(...) list in the WHERE condition on an NDB table could cause mysqld to crash. (Bug #29185)

Memory corruption could occur due to a problem in the DBTUP kernel block. (Bug #29229)

A race condition could result when nonmaster nodes (in addition to the master node) tried to update active status due to a local checkpoint (that is, between NODE_FAILREP and COPY_GCIREQ events). Now only the master updates the active status. (Bug #28717)

An invalid comparison made during REDO validation that could lead to an Error while reading REDO log condition. (Bug #29118)

UPDATE IGNORE statements involving the primary keys of multiple tables could result in data corruption. (Bug #28719)

The management client's response to START BACKUP WAIT COMPLETED did not include the backup ID. (Bug #27640)

Setting InitialNoOfOpenFiles equal to MaxNoOfOpenFiles caused an error. This was due to the fact that the actual value of MaxNoOfOpenFiles as used by the cluster was offset by 1 from the value set in config.ini. (Bug #28749)

Disk Data: When loading data into a cluster following a version upgrade, the data nodes could forcibly shut down due to page and buffer management failures (that is, ndbrequire failures in PGMAN). (Bug #28525)

Disk Data: When dropping a page, the stack's bottom entry could sometime be left “cold” rather than “hot”, violating the rules for stack pruning. (Bug #29176)

Disk Data: Repeated INSERT and DELETE operations on a Disk Data table having one or more large VARCHAR columns could cause data nodes to fail. (Bug #20612)

Cluster API: An invalid error code could be set on transaction objects by BLOB handling code. (Bug #28724)

Cluster API: The timeout set using the MGM API ndb_mgm_set_timeout() function was incorrectly interpreted as seconds rather than as milliseconds. (Bug #29063)

Added the mysqld option --ndb-cluster-connection-pool that enables a single MySQL server to use multiple connections to the cluster. This enables scaling out using multiple MySQL clients per SQL node instead of or in addition to using multiple SQL nodes with the cluster.

For more information about this option, see MySQL Server Options and Variables for MySQL Cluster.

New cluster management client DUMP commands were added to aid in tracking transactions, scan operations, and locks. See DUMP 2350, DUMP 2352, and DUMP 2550, for more information.

Multiple operations involving deletes followed by reads were not handled correctly.

(Bug #28276)

Cluster API: Using NdbBlob::writeData() to write data in the middle of an existing blob value (that is, updating the value) could overwrite some data past the end of the data to be changed. (Bug #27018)

An --ndb-wait-connected option has been added for mysqld. When used, it causes mysqld wait a specified amount of time to be connected to the cluster before accepting client connections.

Cluster API: The Ndb::startTransaction() method now provides an alternative interface for starting a transaction.

Cluster API: It is now possible to iterate over all existing NDB objects using three new methods of the Ndb_cluster_connection class: