Re: Recovery after node restart

[Annie Blomqvist]

I did another test where the node was down for a longer time period.
It actually created a new router eventually, but event then the cluster this not recover properly.

NAME READY STATUS RESTARTS AGE
pod/mysql-cluster-0 0/2 Terminating 14 6d6h
pod/mysql-cluster-1 2/2 Running 14 (100m ago) 6d6h
pod/mysql-cluster-2 2/2 Running 14 (128m ago) 6d6h
pod/mysql-cluster-router-7988469f76-8kkwp 1/1 Running 0 84m
pod/mysql-cluster-router-7988469f76-fcbff 0/1 Terminating 1294 6d6h

The mysql and the router from the failing node seem to be stuck in a terminating state. This also means that the cluster is now down to 2/3 replicas...

Would like to know if there is any way of changing the setup so that:

1. a short down time for a node does not result in a broken cluster/router does no longer seem to be a member of the cluster.
2. a longer down time does not result in a partly down cluster due to cluster and router to be stuck in terminating state.