1Z0-908 practice question 19 of 25

An attempt to recover an InnoDB Cluster fails. Examine this set of messages and responses: host3:3377 ssl JS > dba.rebootClusterFromCompleteOutage()…

Full set: 140 questions
An attempt to recover an InnoDB Cluster fails. Examine this set of messages and responses: host3:3377 ssl JS > dba.rebootClusterFromCompleteOutage() Reconfiguring the default cluster from complete outage… The instance ‘host1:3377'’ was part of the cluster configuration. Would you like to rejoin it to the cluster? [y/N]: y The instance ‘host2:3377’ was part of the cluster configuration. Would you like to rejoin it to the cluster? [y/N]: y Dba.rebootClusterFromCompleteOutage: The active session instance isn’t the most updated in comparison with the ONLINE instances of the Cluster’s metadat a. Please use the most up to date instance: ‘host1:3377’. (RuntimeError) Which statement is true?
Answer options
A: The instance deployed on host3 must be rebuilt with a backup from the primary instance.
B: The cluster is running and there is at least one ONLINE instance.
C: The instance deployed on host3 must be synchronized from a donor deployed on host1 by using the command cluster.addInstance(‘host1:3377’).
D: It is possible to determine the most up-to-date instance by comparing different global transaction identifier (GTID) sets with GTID_SUBSET(set1,set2). Correct answer
E: The active session instance is invalid and must be re-created by using the command shell.connect (‘host3:3377’).
Correct Answer:
D. It is possible to determine the most up-to-date instance by comparing different global transaction identifier (GTID) sets with GTID_SUBSET(set1,set2).
Explanation

The error message The active session instance isn’t the most updated... Please use the most up to date instance: ‘host1:3377’ occurs because the dba.rebootClusterFromCompleteOutage() command must be executed from the cluster member that processed the most transactions before the outage. InnoDB Cluster determines this by comparing the Global Transaction Identifier (GTID) sets of all member instances. The instance with the superset of all GTIDs is considered the most up-to-date. The error explicitly states that host1:3377 is this instance. Therefore, the statement that this determination is made by comparing GTID sets is the most accurate technical explanation for the situation presented.

Why Incorrect

A. Rebuilding from a backup is a manual recovery method, not the standard procedure required by InnoDB Cluster tooling, which is designed to handle resynchronization automatically.
B. The command being executed, dba.rebootClusterFromCompleteOutage(), explicitly indicates that the cluster is not running and is in a state of complete outage.
C. The cluster.addInstance() command is used to add a new instance to an already running and healthy cluster, not to rejoin an existing member during a recovery from an outage.
E. The session to host3:3377 is valid, but the instance itself is not the most up-to-date. Reconnecting to the same outdated instance will result in the same error.

References

1. MySQL 8.0 Reference Manual, Section 21.5.3.2, "Rebooting a Cluster from a Complete Outage": The documentation states, "It then tries to discover the other members of the cluster and checks which of them has the most recent set of transactions. To do this, it compares the GTID sets of the members. If the instance you are connected to is not the one with the most recent transactions, the operation stops and returns an error." This directly supports that GTID set comparison is the underlying mechanism.

2. MySQL 8.0 Reference Manual, Section 17.1.3.6, "GTID Set Functions": This section describes the GTIDSUBSET() function, which "returns true if all GTIDs in set1 are also in set2." This is the specific function used to perform the comparison of GTID sets to find the most up-to-date member.

3. MySQL 8.0 Reference Manual, Section 21.4.3.1, "Adding an Instance to a Cluster": The documentation for cluster.addInstance(instance) confirms its purpose is to "add a MySQL server instance to a cluster," which is not the operation required in a complete outage recovery scenario for an existing member. This refutes the action proposed in option C.

Scroll to Top