18.6.6.4 Exit Action

If READ_ONLY is the exit action, the instance switches MySQL to super read only mode by setting the system variable super_read_only to ON. When the member is in super read only mode, clients cannot make any updates, even if they have the SUPER privilege. However, clients can still read data, and because updates are no longer being made, there is a probability of stale reads which increases over time. With this setting, you therefore need to pro-actively monitor the servers for failures. This exit action is the default from MySQL 8.0.15. After this exit action is taken, the member's status is displayed as ERROR in the view of the group.

If OFFLINE_MODE is the exit action, the instance switches MySQL to offline mode by setting the system variable offline_mode to ON. When the member is in offline mode, connected client users are disconnected on their next request and connections are no longer accepted, with the exception of client users that have the CONNECTION_ADMIN or SUPER privilege. Group Replication also sets the system variable super_read_only to ON, so clients cannot make any updates, even if they have connected with the SUPER privilege. This exit action prevents both updates and stale reads (with the exception of reads by client users with the stated privileges), and enables proxy tools such as MySQL Router to recognize that the server is unavailable and redirect client connections. It also leaves the instance running so that an administrator can attempt to resolve the issue without shutting down MySQL. This exit action is available from MySQL 8.0.18. After this exit action is taken, the member's status is displayed as ERROR in the view of the group (not OFFLINE, which means a member has Group Replication functionality available but does not currently belong to a group).

If ABORT_SERVER is the exit action, the instance shuts down MySQL. Instructing the member to shut itself down prevents all stale reads and client updates, but it means that the MySQL Server instance is unavailable and must be restarted, even if the issue could have been resolved without that step. This exit action was the default from MySQL 8.0.12, when the system variable was added, to MySQL 8.0.15 inclusive. After this exit action is taken, the member is removed from the listing of servers in the view of the group.

Applier error - There is an error in the replication applier. This issue is not recoverable.

Distributed recovery not possible - There is an issue that means Group Replication's distributed recovery process (which uses remote cloning operations and state transfer from the binary log) cannot be completed. Group Replication retries distributed recovery automatically where this makes sense, but stops if there are no more options to complete the process. For details, see Section 18.4.3.3, “Fault Tolerance for Distributed Recovery”.

Group configuration change error - An error occurred during a group-wide configuration change carried out using a UDF, as described in Section 18.4.1, “Configuring an Online Group”.

Primary election error - An error occurred during election of a new primary member for a group in single-primary mode, as described in Section 18.1.3.1, “Single-Primary Mode”.

Unreachable majority timeout - The member has lost contact with a majority of the group members so is in a minority, and a timeout that was set by the group_replication_unreachable_majority_timeout system variable has expired.

Member expelled from group - A suspicion has been raised on the member, and any timeout set by the group_replication_member_expel_timeout system variable has expired, and the member has resumed communication with the group and found that it has been expelled.

Out of auto-rejoin attempts - The group_replication_autorejoin_tries system variable was set to specify a number of auto-rejoin attempts after a loss of majority or expulsion, and the member completed this number of attempts without success.