In our lab environment, we received some emails from our provider, DigitalOcean saying a few nodes would be going down, because of issues with the physical nodes. Since we run a 9-node Galera Cluster in the lab, across 3 regions (San Francisco, London and Singapore), this posed an interesting problem!
Upon connecting to one of the nodes, we see:
mysql> show status like 'wsrep_cluster_size'; +--------------------+-------+ | Variable_name | Value | +--------------------+-------+ | wsrep_cluster_size | 7 | +--------------------+-------+ 1 row in set (0.02 sec)
So we know that 2 of the nodes are down. Presuming you do not have any monitoring setup (might we recommend …
[Read more]