Hello!
I'm running an ESXi host with 5.0 update 02 from end of december 2012, build 914586.
I'm using an LSI RAID controller with 4x1TB SAS hard drives in a RAID5 array, splitted in two logical drive, 100GB for System and 2900GB for Data.
The ESXi datastore with the virtual machines is located on a MBR/VMFS partition on the 100GB logical drive.
The 2900GB logical drive is formated with GPT/NTFS (native) and is used as a independent, permanent RDM passthrough disk in a virtual machine.
Now, I tried to update the host to the latest patch ESXi500-201303001, build 1024429 from March 2013 (in the hope, that RDM passthrough of a 3TB SATA harddrive, connected to the Intel onboard SATA controller, works now with correct disk size).
The patch installed successfully, but after reboot, powering on the virtual machine with the RDM passthrough'ed logical LSI disk failed.
A strange error occurs; the vmdk of this disk could not be opened, and, if I remember correctly, something about snapshots - even with an independent, permanent disk, making no use of snapshots, I thought.
After this error, I rolled back to the old build 914586 and the virtual machine works again fine....
-> Could it be, that RDM passthrough of local hard drives is generally broken in the latest build from March 2013 - does anybody have similar experience?
Thanks.