Removing Writer/Reader from instance node name #22
Conversation
LewS
left a comment
There was a problem hiding this comment.
Avoids the confusion of misnamed writer/reader instances, but there is also the r53 records, which I'm less sure that we need. The problem with renaming these resources is that it will create new resources for any stack not using a locked version of the component. I think we should consider the creation of a couple of conditions, so that the old behavior defaults and the new behavior can be chosen.
There was a problem hiding this comment.
Avoids the confusion of misnamed writer/reader instances, but there is also the r53 records, which I'm less sure that we need. The problem with renaming these resources is that it will create new resources for any stack not using a locked version of the component. I think we should consider the creation of a couple of conditions, so that the old behavior defaults and the new behavior can be chosen.
There was a problem hiding this comment.
Yeah, it's going to add nodes to the cluster, there is going to failover at some point. Route53 is pointing to the cluster-endpoint, or at least should be. Let's see if others agree on the naming them we can have another thread on how to rollout
There was a problem hiding this comment.
I can't think of anything better for the naming convention, my only thought was something like DBClusterInstancePeer for the reader(s), suggesting that they're the same (though maybe it would be too misleading). Clone? Twin?

No description provided.