Closed misterbisson closed 7 years ago
I've now confirmed this is not just a theoretical issue, but a real bug. After scaling down from four MySQL instances to three, where the one instance Docker Compose chose to kill off was the MySQL primary, WordPress didn't detect the replacement primary.
The backends stanza in WordPress'
containerpilot.json
tracksmysql
,nfs
, andmemcached
, but doesn't trackmysql-primary
. This has not been a problem so far because electing a newmysql-primary
has so far always changed the pool ormysql
instances. That assumption has been safe so far, but it doesn't seem like one we need to enshrine here.