MySQL InnoDB Cluster not healthy: None

I used to have a functional 3 node mysql-innodb-cluster but I get one unit removed and now I´ve two units (one RW and another RO). I´m trying to add a new unit to get 2 RO but the deployment stuck with “MySQL InnoDB Cluster not healthy: None”. Just on this Unit. The other two keeps RW and RO. Is there anything that I´m missing on deployment or configuration?

This is the command that I´ve used to deploy:

juju add-unit --to lxd:1 mysql-innodb-cluster

@rafael.cardoso interested to know if you resolved this. I find something similar myself, in my case I am trying add a new instance before deleting an existing one.

In your case, check the health of the cluster using

juju run-action mysql-innodb-cluster/leader --wait cluster-status

You will probably find the artefact of the instance you removed present. This will certainly cause problems if it has the same IP as the new instance you are adding. You could try removing that;

juju run-action mysql-innodb-cluster/leader remove-instance address="a.b.c.d:3306" force=true

Hi all,

Thank you for bringing this matter to our attention.

It sounds like you’ve found a bug in the charm! Bugs for the mysql-innodb cluster charm live at Bugs : charm-mysql-innodb-cluster. If there isn’t an existing bug open about the issue, please file one.

(If the bug turns out to be in Juju rather than the charm, we can transfer the bug to the juju project from there.)

Thank you!
~ PeteVG

I raised this bug.

Hi @rafael.cardoso,

It requires 3 active units to get the cluster running.

Out of any scenario - you can always reset the mysql-innodb-cluster using the below command
juju run-action mysql-innodb-cluster/leader reboot-cluster-from-complete-outage --wait

To check the status of the cluster (after the reset is complete)
juju status mysql-innodb-cluster

Thanks.