How does solaris work out what virtual disk geometry it will use for a LUN?
We are moving from one storage subsystem to an EMC vmax. To migrate the data we are using host based mirroring since the luns are mirrored using SLVM to support our multiple sites and cluster config.
The LUNs that are presented from the VMAX report in luxadm display as being larger than the LUNs we are replacing, but by the time we put them into SLVM and a couple of cylinders are reserved for state databases, the useable data space is less than the data partition on the old lun.
What I'm trying to understand is the algorithm that solaris uses to define the disk geometry that it overlays onto a LUN. I know that there's a limit of 65533 usable cylinder (65535 total) for a lun.