MySQL 8.4 and MariaDB 11.8 testing

[dveeden]

More testing is needed for MySQL 8.4 and MariaDB 11.8. These are the latest LTS versions. We could also add the latest non-LTS versions if needed.

For MySQL 8.4:

• binlog_checksum=NONE/CRC32
• binlog_format=row/statement/mixed
• binlog_row_image=minimal
• binlog_row_value_options=PARTIAL_JSON
• binlog_transaction_compression=ON
• binlog_rows_query_log_events=ON
• binlog_encryption=ON (probably doesn't affect us)
• Tagged GTIDs
• Annotated row events

For MariaDB 11.8:

• binlog_format
• binlog_checksum
• binlog_legacy_event_pos (new in 11.4) #12416
• binlog_row_image
• binlog_row_metadata
• Replication with account with minimal privileges (After DM: Improve MariaDB permission handling #12404 has been merged)
• Annotated row events

Optional extras

• RDS MariaDB
• RDS MySQL

For binlog compression in MariaDB and encryption in both MySQL and MariaDB we are probably not affected as this is transparent to the network protocol if I understand this correctly.

Expected outcome for each item: Either supporting the feature or have a reasonable error.

Findings so far

• binlog_legacy_event_pos probably needs to be disabled until it is fully supported. For this the check needs to be updated to check for this variable and also the documentation needs to include this.
• DM: MariaDB 11.8 DPANIC due to GTID set update failure #12423
• DM: MariaDB and MySQL specific DDL options cause failures #12429

[dveeden]

/assign @dveeden

[dveeden]

/label MySQL
/label MariaDB
/area DM

[ti-chi-bot]

@dveeden: The label(s) mysql cannot be applied. These labels are supported: MariaDB, affects-6.5, affects-7.1, affects-7.5, affects-8.1, affects-8.5, affects-9.0, bug-from-internal-test, bug-from-user, duplicate, may-affects-6.5, may-affects-7.1, may-affects-7.5, may-affects-8.1, may-affects-8.5, needs-cherry-pick-release-6.5, needs-cherry-pick-release-7.1, needs-cherry-pick-release-7.5, needs-cherry-pick-release-8.1, needs-cherry-pick-release-8.5, needs-cherry-pick-release-9.0-beta.2, nextgen, question, release-blocker, wontfix.

Details

In response to this:

/label MySQL
/label MariaDB
/area DM

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository.

[dveeden]

/label compatibility

[ti-chi-bot]

@dveeden: The label(s) compatibility cannot be applied. These labels are supported: MariaDB, affects-6.5, affects-7.1, affects-7.5, affects-8.1, affects-8.5, affects-9.0, bug-from-internal-test, bug-from-user, duplicate, may-affects-6.5, may-affects-7.1, may-affects-7.5, may-affects-8.1, may-affects-8.5, needs-cherry-pick-release-6.5, needs-cherry-pick-release-7.1, needs-cherry-pick-release-7.5, needs-cherry-pick-release-8.1, needs-cherry-pick-release-8.5, needs-cherry-pick-release-9.0-beta.2, nextgen, question, release-blocker, wontfix.

Details

In response to this:

/label compatibility

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository.

[dveeden]

From mariadb-binlog mariadb-log.000002 with binlog_legacy_event_pos=OFF (default):

...
# at 709
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0x03600827 	Annotate_rows:
#Q> insert into t1 values (1),(2),(3)
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0xe955ff59 	Table_map: `test`.`t1` mapped to number 18
# at 754
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0x895ab893 	Write_rows: table id 18 flags: STMT_END_F
...

Note how the end_log_pos is 0 for these events.

With a --hexdump option added you can see that Master Pos is 00 00 00 00, but Size is filled in.

# at 653
# at 709
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0x03600827 
# Position
#           |Timestamp   |Type |Master ID   |Size        |Master Pos  |Flags
#      2c5  |55 00 26 69 |a0   |01 00 00 00 |38 00 00 00 |00 00 00 00 |00 00
#
#      2d8  69 6e 73 65 72 74 20 69  6e 74 6f 20 74 31 20 76  |insert into t1 v|
#      2e8  61 6c 75 65 73 20 28 31  29 2c 28 32 29 2c 28 33  |alues (1),(2),(3|
#      2f8  29 27 08 60 03                                    |)'.`.|
#
# Event: 	Annotate_rows:
#Q> insert into t1 values (1),(2),(3)
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0xe955ff59 
# Position
#           |Timestamp   |Type |Master ID   |Size        |Master Pos  |Flags
#      2c5  |55 00 26 69 |13   |01 00 00 00 |2d 00 00 00 |00 00 00 00 |00 00
#
#      2d8  12 00 00 00 00 00 01 00  04 74 65 73 74 00 02 74  |.........test..t|
#      2e8  31 00 01 03 00 00 59 ff  55 e9                    |1.....Y.U.|
#
# Event: 	Table_map: `test`.`t1` mapped to number 18
# at 754
#251125 19:15:33 server id 1  end_log_pos 0 CRC32 0x895ab893 
# Position
#           |Timestamp   |Type |Master ID   |Size        |Master Pos  |Flags
#      2f2  |55 00 26 69 |17   |01 00 00 00 |30 00 00 00 |00 00 00 00 |00 00
#
#      305  12 00 00 00 00 00 01 00  01 01 fe 01 00 00 00 fe  |................|
#      315  02 00 00 00 fe 03 00 00  00 93 b8 5a 89           |...........Z.|
#
# Event: 	Write_rows: table id 18 flags: STMT_END_F

[dveeden]

cc @lance6716 @alastori @OliverS929

[dveeden]

For binlog_format and binlog_row_image we have checks already. No need to check them again.