Mixed mode VMDKs

[continuum]

For advanced users only ....
Did you know that you can create vmdks with preallocated and sparse areas ?
I often use dd images of VMFS-metadata.
To create such a dump I do
dd if=/dev/disks/naa******:1 bs=1M count=1536 of=vmfs.meta
To edit such a dump with an ESXi VM you can then create a vmdk like this:

# Disk DescriptorFile

version=1

encoding="windows-1252"

CID=2e012c60

parentCID=ffffffff

isNativeSnapshot="no"

createType="twoGbMaxExtentSparse"

# Extent description

RW 2048 FLAT firstmb 0

RW 3145728 FLAT "vmfs.meta" 0

RW 4261412864 SPARSE "8tb-vmfs-s001.vmdk"

RW 4261412864 SPARSE "8tb-vmfs-s002.vmdk"

RW 4261412864 SPARSE "8tb-vmfs-s003.vmdk"

RW 4261412864 SPARSE "8tb-vmfs-s004.vmdk"

RW 111149056 SPARSE "8tb-vmfs-s005.vmdk"

RW 4261412864 SPARSE "8tb-vmfs-s006.vmdk"

RW 2048 FLAT lastmb 0

# The Disk Data Base

#DDB

The example vmdk is about 10TB in size and uses a vmfs-meta dump from a normal harddisk or LUN dump.
NOTE: you do NOT need to specify the usual geometry-parameters.
NOTE: to expand such a vmdk do NOT use the  vmware-vdiskmanager -x function

- instead expand the sparse area  by copying one of the sparse-slices and adjusting the descriptor yourself.
Possible use case:
extract vmfs-meta dump at a customer.
Fix errors in the dump inside a ESXi-VM.
Reinject modified dump at customer .....
DISCLAIMER: be careful !!!

Interested ? - lets chat via skype "sanbarrow"