Comment 9 for bug 1881735

Revision history for this message
Sérgio Manso (sergiomanso) wrote :

I've also faced this issue in field with cs:mysql-innodb-cluster-3

The cluster could not be initialized because one of the units got stuck in group_replication statement with the exact same message.

Workaround: Get the mysql password, stop group replication in the failed unit and add it again to the cluster.

1. juju run --unit mysql-innodb-cluster/<leader> leader-get
2. juju ssh <stucked_unit>
3. mysql -u root -p
4. [mysql]> stop group_replication;
5. juju run-action mysql-innodb-cluster/<leader> add-instance address="<stucked_unit_ip>"