That's the behaviour of the vmkfstools tool of VMWarethat GhettoVCBis using to create the clone. You can run 'vmkfstools --punchzero disk.vmdk' afterwards but it would be nice if this was a build in option in GhettoVCB just like the compression option.
Does anyone got fix for it?
I got the script working for me, backup and restore both are working.
But the backup creates entire size of thin VM on destination datastore.
As I understand it, when writing to VMDKs over NFS, ghettoVCB — and, more generally, ESXi — only writes to non-zero blocks in the destination file; some NFS servers allocate space on the target filesystem for the unwritten blocks, while others do not.
When investigating NFS server options, the key term you need to look for is "sparse file" support, and note that the NFS server software and the underlying filesystem it serves must both support sparse files to enable "thin" backups with ghettoVCB.
Also, be aware that, even on systems that support sparse files, the file size reported in directory listings for a sparse file typically reflects the "full" size, including unwritten blocks, rather than the actual amount of disk space used. On UNIX-like systems, the du command typically reports the actual disk space used.