>> Most of
>> the time, there should be 2 identical nodes for root failover, does
>> this solution still work for those none-identical nodes for root
>I don't understand your question. Can you explain?
For example, if there is a two-nodes root failover cluster where one node with SCSI local disk and another node with IDE local hardisk. Is it OK to just copy contents of one of the nodes local boot partition to /boot directory?