All COE-based products, such as CAS, Management Center or Reporter 10.x and later versions may hit this issue when the storage is SAN based. We have so far seen this issue with storage made by Netapp and Pure.
COE based products use EFI firmware. The size of the IO being passed is overflowing the queue on the storage system, leading to latency. As a result when booting up, the VM fails to find the necessary disk files needed to boot the image.
This issue is well documented here with EMC storage:
Change the Disk.DiskMaxIOSize setting of the ESX host as per this article:
This is a global setting and will affect all VM's configured on the ESX host. You may well need to reload the OVF file after changing the Disk.DiskMaxIOSize setting.
The following is included in the Pure storage best practice guide:
Disk.DiskMaxIOSize 4096 32768
Dictates the largest IO the ESXi kernel will transmit. This must be lowered when using EFI-enabled virtual machines.