First >> the file limit on VMFS 5 is > 100,000 on a datastore. Does this mean vm files like .log // flat file // swap file etc. These type of files cannot be more than 100,000 ?
OR if we create a windows machine and its files which are on vmdk -- are they also counted in this 100,000 ?
This relates only to files on the VMFS volume (like vmx, vmdk, swap, vmware.log, HA-datastore heartbeating, dvS config etc.) itself and not to files within the GuestOS filesystem on a virtual disk, which ESXi is not aware of in the first place. You won't get anywhere near this limit unless you run literally thousands of VMs on a single VMFS datastore (and something must be wrong if anyone is doing that).
Second >> As per VMware KB VMware KB: Frequently Asked Questions on VMware vSphere 5.x for VMFS-5
VMFS 5 Uses SCSI_READ16 and SCSI_WRITE16 cmds for I/O (VMFS-3 used SCSI_READ10 and SCSI_WRITE10 cmds for I/O).
What is the technical difference scsi_read_10 and scsi_write_16 and same for write ?
This refers to low-level details of the SCSI protocol standard. I don't think this is anything we really need to concern ourselves with, but if you're interested, check SCSI protocol documentation like this document:
http://www.seagate.com/staticfiles/support/disc/manuals/scsi/100293068a.pdf
The various READ (X) commands are explained from chapter 3.19 onwards while the WRITE (X) commands are defined from 3.53.
In a nutshell, it seems like the only major difference is that the (16) protocol commands leave more space for the block addressing and transfer length header fields. This is probably one of the things that enables VMFS5 to address block storage devices (aka LUNs) larger than 2TB as opposed to VMFS3.
--
http://alpacapowered.wordpress.com
