There is no simple answer, the usual way to deal with this case is to automate the full stack, so in case of such changes, destroy and rebuild the machine from scratch to avoid old confs.
Another way is to cancel the blur time in the cookbook (one or two releases)
To talk about recent events, let's say you added an attribute new_mountpointto the old one mountpointand updated your recipe with the new mountressource using the new attribute and updated the old mountresource to remove the old mount point.
The next change, new_mountchanges for the new path, and mountpointtakes the previous valuenew_mount
source
share