fix #4869

when delete node via seaos apply -f Clusterfile, the target node will not exist in the Clusterfile, in that case, if the node has a different passwd with other nodes, the scale down process will failed.

my solution is passing the 'current' clusterfile when NewScaleProcessor, it will not affect the sealos add and sealos delete because the add and delete processes pass current clusterfile originally. It will only affect the sealos apply when it comes scale process.