Currently we can't differentiate which EBS volumes are regular, full-time ones and which are daytime ones. This could cause problems if a deployment picked up daytime ones with slightly old data.
While at it, instead of using an attribute named logsearch for allocation awareness, generalize it and call it replica_role.
These changes allow us to be a bit more flexible in how many instances we want to use and potentially having more than two replication levels.
Currently we can't differentiate which EBS volumes are regular, full-time ones and which are daytime ones. This could cause problems if a deployment picked up daytime ones with slightly old data.
While at it, instead of using an attribute named
logsearchfor allocation awareness, generalize it and call itreplica_role.These changes allow us to be a bit more flexible in how many instances we want to use and potentially having more than two replication levels.