On Dec 9, 2007 11:08 PM, Alan Altmark <[EMAIL PROTECTED]> wrote: > Are you really suggesting inserting CP into the middle of SCSI I/O in > order to create FC minidisks? If so, that requirement needs to be made > with bold underscored italics to get attention. The current thinking is > that guests need to be managed just like their distributed cousins when it > comes to SAN connectivity.
You make it sound like a virtualization is some novel idea that we would not be certain if it would stay around. CP has the ability to isolate the virtual machine from details (like where the data is, what type of device the data is, etc). When the guest does not need to be aware of such details, CP has the freedom to arrange things has one sees fit on a system wide level (duplication of data, multiple paths, amount of cache, etc). As long as Linux configuration needs to have intimate knowledge about how its data is managed, things like performance management and disaster fall-back will remain very fragile at least. Given the way Linux uses disk space, I would suggest that z/VM should provide the virtual machine with a bunch of blocks (probably over-commit that as well, like with SFS). A high-level interface for the I/O operation will give CP freedom in performing the operations in a way that is efficient (with the added bonus that you don't need to virtual machine to be involved to drive the actual I/O). Rob -- Rob van der Heij Velocity Software, Inc http://velocitysoftware.com/
