Hi, i'm using Copycat version 1.2.3 and i've got a question related to membership changes. when i'm removing a member from the cluster, remove its state from disk and bring the node back online, the node doesn't catch up with its peers in my test. is this the expected behaviour? i'd expect the node to catch up by replaying and applying the log entries to the state machine. i'm not compacting the logs and i'm not using snapshotting. thanks.