When an operation is not recorded in the binlog it doesn‘t have gtid assigned.
In this example we are disabling the recording of the events before.

Data dictionary tables are stored within mysql schema.
INFORMATION_SCHEMA offers views over data dictionary tables for manual access but they are not part of the data dictionary.
mysql.users is part of the Grant System tables but not of the data dictionary.

Query Cache was deprecated in MySQL 5.7 and completely removed in 8.0. You cannot use it anymore.

Correct:

A. db

The db table is a core part of the Grant system tables, located in the mysql schema. It stores privileges that are granted to users for specific databases (schemas). A row in the db table grants a set of privileges to a user connecting from a specific host for a specific database.

C. user

The user table is the most fundamental Grant system table. Also located in the mysql schema, it stores the list of all user accounts, their authentication information (like password hashes), global privileges, and the connection-related restrictions applied to them.

D. password_history

The password_history table is a security-related table within the Grant system (in the mysql schema). It is used to implement the password re-use policy enforced by the validate_password component. It stores a history of the hash values of a user‘s previous passwords to prevent them from reusing recent passwords.

Incorrect:

B. host

The host table was a legacy Grant table in older versions of MySQL (prior to 8.0) that stored host-specific privileges for anonymous users. In MySQL 8.0, the functionality of the host table has been integrated into the db table, or replaced by modern, more granular privilege controls using the user table and explicit host matching in grants. The host table is not actively used as part of the core Grant system for defining privileges in modern MySQL 8.0 deployments.

In Group Replication Single-Primary mode, only 1 server accepts writes simultaneously.

Roles can be not enabled by default. Users can enable them individually or all excluding whatever is not needed.

/*+ table index1, index2 */ is the correct syntax

A new InnoDB tablespace stores the data dictionary: mysql.ibd under the datadir

MySQL Enterprise Firewall can run in 3 different modes. Initially you will run in Recording until you have captured most of the normal activity:
Recording: the whitelist is populated with the new queries and those are executed
Detecting: queries not present in the whitelist are executed but logged in the Firewall log
Protecting: queries not present in the whitelist are blocked and logged in the Firewall log

A page is inserted at the head of the Old Sublist

OLD and NEW are XML formats.
OLD has the information defined as attributes
NEW has the information defined as sub-nodes

Correct:

C. Connection events for any user

The set of three commands configures MySQL Enterprise Audit to log connection events for every user who attempts to connect to the server.

The first command defines the filter variable @f to include only events belonging to the “connection“ event class: ““class““: { ““name““: ““connection““ }.

The second command registers this filter definition under the name ‘log_filter‘.

The third command applies this ‘log_filter‘ to the user identifier ‘%‘. In MySQL access control, the wildcard character ‘%‘ matches all users (both local and remote) and all hostnames. Therefore, the audit log will capture connection events for any user trying to connect.

Incorrect:

A. Connection events for remote users

While connection events for remote users are included, the ‘%‘ wildcard in the user specification (audit_log_filter_set_user(‘%‘, ‘log_filter‘)) applies the filter to all users, including users connecting locally (e.g., from localhost or via a Unix socket). This option is too restrictive.

B. Everything for any user

The filter definition explicitly restricts logging to only the “connection“ event class: ““class““: { ““name““: ““connection““ }. The connection class does not include other events like query, table, or stored_procedure events, so the log will not contain “Everything.“

D. Everything for remote users

This option is incorrect on two counts:

“Everything“ is wrong because the filter explicitly limits the logging to only the “connection“ class.

“remote users“ is wrong because the wildcard ‘%‘ applies to all users, including local users.

The maximum size for string columns is always given on bytes as different character sets can require different amount of bytes per character

InnoDB page size cannot be changed without reinitializing the database

SYSTEM_VARIABLES_ADMIN enables server configuration changes.
SESSION_VARIABLES_ADMIN enables setting restricted session system variables.
SUPER_ADMIN doesn‘t exist

The official documentation states that this will crash the server and corrupt the backup.

Read-ahead operations don‘t promote pages; only when a page is directly read is promoted in the buffer pool

You can remove column histograms using ANALYZE TABLE… DROP HISTOGRAM ON… syntax.
ANALYZE TABLE will not remove the existing histograms.

Prior MySQL 8.0 plugins could be written in C or C++. With MySQL 8.0 only C++ is a valid choice.

All of the above.
General tablespaces is a new feature of MySQL 8.0. They allow to store multiple tables within the same tablespace but outside system tablespaces

MySQL 8.0.17 performs the operations that were previously done by mysql_upgrade automatically

JSON_STORAGE_SIZE function returns the space used by the binary representation of a JSON.
Before MySQL 8.0 it was not possible to obtain this value and you had to approximate it using LENGTH

Only users with the privilege SERVICE_CONNECTION_ADMIN can connect through the new interface enabled with admin_address and admin_port

mysql_upgrade has been deprecated. The data dictionary tables are automatically upgraded when the mysqld process starts and detects that the version of the dictionary is older than the current binary.

Group Replication requires that the majority of the servers are available after a failure.
In this case, only with 5 or 6 servers we would have more than half servers available if 2 failed.

The attribute WIDTH does not affect the space required to store a numeric value.
Be aware that WIDTH has been deprecated as of MySQL 8.0.17 as it doesn‘t affect in any way NUMERIC types

mandatory_roles defines a list of roles that will be assigned to all the users without explicitly granting them. They cannot be revoked or dropped.

In MySQL 8.0 after enabling skip_name_resolve only users defined with IP addresses will be considered.
The user will fail to connect as it was created using the hostname.

In MySQL 8.0 metadata files are not used and the dictionary has been moved to transactional tables.
Different storage engines can also store information in their individual dictionary data.

caching_sha2_password is the new default authentication plugin for clients in MySQL 8.0

Only accounts with SYSTEM_USER can modify other system accounts.
If the user has all privileges with the re-grant option it will able to modify other user accounts.

With a fill factor of 100, InnoDB will leave 1/16 of the space in clustered index pages free for future index growth

Log tables continue to use the CSV engine

MySQL 8.0 introduces INVISIBLE indexes; indexes marked as invisible are maintained but the optimizer won‘t use them.
Making an index INVISIBLE/VISIBLE is a fast, in-place operation.

The WIDTH attribute for NUMERIC columns doesn‘t constrain the range of values that can be stored or displayed. It has been deprecated in MySQL 8.0.17 and you should avoid it.

Each server in a multi-source replication must have a unique server_id.

The source doesn‘t perform the commit until at least 1 replica has acknowledged that the event is stored in its relay log.

Invisible indexes are ignored by the optimizar even if they are indicated in a hint. Modifying the optimizer_switch at session level is the less disruptive method.

The primary has all the required binlogs for the replication to continue. This can be identified because the primary gtid_purged is smaller than the replica gtid_executed.
In this scenario if the replication will resume without any additional intervention.

Full backup contains a complete copy of the data.
Incremental backup contains the data modified since the last backup. The last backup could be another incremental backup.
Differential backup contains the data modified since the last full backup.

1. Restore the full backup, this will be your base
2. Restore the incremental backup, this will be applied over your base

Replication SQL thread applies events to the replica data, slave_parallel_workers modify the number of threads applying those events.

TIME and TIME(0) are equivalents

Correct:

B. R-TREE

The InnoDB storage engine, which is the default for MySQL 8.0, uses the R-Tree data structure specifically for storing spatial indexes (defined with the SPATIAL INDEX keyword).

R-Trees are specialized, multi-dimensional data structures designed to efficiently index non-point objects (like lines, polygons, and minimum bounding rectangles—MBRs) and perform spatial queries such as range searches (e.g., “find all objects in this area“). This structure is necessary because traditional B-Trees are optimized only for one-dimensional data.

Incorrect:

A. DOUBLE LINKED LIST

A double linked list is a linear data structure used for maintaining ordered sequences (like the records within an index page), but it is not used as the fundamental tree structure for any type of index in InnoDB.

C. GRAPH

A graph is a network model of data relationships. While specialized graph databases exist, InnoDB does not use a general graph structure to store its indexes. The index structures it uses are strictly tree-based (B-Tree and R-Tree).
D. B-TREE

The B-Tree (specifically a B+ Tree) is the default and primary data structure used by InnoDB for all non-spatial indexes (primary keys, unique indexes, and secondary indexes). B-Trees are optimized for one-dimensional data ordering and range lookups. However, because spatial data is multi-dimensional (e.g., X and Y coordinates), a standard B-Tree is inefficient for spatial queries, which is why InnoDB switches to the R-Tree structure for SPATIAL INDEX.