The latest version of MySQL 8.0.20 is officially released. Like the previous 8.0 series, this release includes bug fixes as well as new features. Let's take a quick look.
Keywords: hash join, InnoDB double write buffer, binary log transaction compression .
-
Added SHOW_ROUTINE permission. Using this authority allows the account to back up stored programs with minimal authority.
-
Improve hash join. Now you can use hash join instead of BNL. Hash join now supports semi-join, anti-join, and outer join.
-
Added new index level optimizer hints. Including: JOIN_INDEX, NO_JOIN_INDEX, GROUP_INDEX, NO_GROUP_INDEX, ORDER_INDEX, NO_ORDER_INDEX, INDEX and NO_INDEX. The new index-level optimizer hints can be used with SELECT, UPDATE and DELETE statements, but the previous ones cannot be used with DELETE.
-
Performance Schema improved. In order to reduce the memory overhead of error detection, Performance Schema gave up collecting session-related error statistics on a global scale. In addition, the error summary reported for each thread, account, user, or host no longer contains global error information.
- SQL syntax improvement. Previously, the INTO clause SELECT of the statement may appear in one of the following two places:
Now, INTO can appear at the end of the SELECT:
In addition, some changes have been made to the INTO for UNION, and the following statements have the same result:
-
X Plugin is improved. A management session is created during startup to obtain the server configuration when the client connection limit is reached. After initialization, you can access MySQL system variables and other bug fixes.
-
The BLOB type sorting changed. The previous non-complete sorting is changed. The server converts the BLOB type into a packaged plug-in for sorting. After testing, the performance has been significantly improved.
-
InnoDB is improved. The storage of the double write buffer is moved from the system table space to the double write file. Moving the double write buffer storage area out of the system table space can reduce write latency, increase throughput, and provide flexibility in the placement of double write buffer pages.
-
Binary log transaction compression. Starting from MySQL 8.0.20, you can enable binary log transaction compression on the MySQL server instance. After enabling binary log transaction compression, the transaction payload will be compressed using the zstd algorithm, and then written as a single event (Transaction_payload_event) to the server's binary log file. When the compressed transaction payload is sent to the slave server, other group replication group members or clients (such as mysqlbinlog) in the replication stream, it remains compressed. They are not decompressed by the receiver thread and are still written to the relay log in their compressed state. Therefore, binary log transaction compression can save both the originator of the transaction and the storage space of the receiver (and its backup), as well as saving network bandwidth when sending transactions between server instances.
-
The new option REQUIRE_TABLE_PRIMARY_KEY_CHECK of the CHANGE MASTER TO statement enables the replicated slave server to choose its own strategy for primary key checking.
- Starting from MySQL 8.0.19, messages sent through the X protocol connection have been compressed. Starting from MySQL 8.0.19, messages sent through the X protocol connection have been compressed. In MySQL 8.0.19, the X protocol uses the library default compression level for each algorithm, and the client cannot negotiate the default compression level. Starting from MySQL 8.0.20, the client can request a specific compression level for the X protocol connection during the negotiation.
The above content is a quick overview of the new features of MySQL 8.0.20. You can visit MySQL's online manual for complete and up-to-date information. https://dev.mysql.com/doc/relnotes/mysql/8.0/en/