On Thu, Jul 27, 2017 at 09:48:44AM -0500, admin at lctn.org wrote: > I'm not sure if this is a vmfs issue or linux, so starting here. > > I have mounted a 3par volume (thinly provisioned vmfs) on my Linux box (64 bit) and I am attempting to copy it to a second hd. I have tried cp, scp, and rsync but the 1.6Tb file ends up with a zero file size on the target hd after many hours of growing in size during the copy. Smaller files copy over fine. It is only the large vmdk with this problem. > > Any ideas how to resolve? The file probably has holes in it. There is no point storing on disk blocks which only contain 0s. So instead, the meta data says which blocks should be all zero. Look at du -h on the file, vs ls -lh. The problem you have is that scp/rsync etc are not preserving the holes. All the empty blocks are being allocated on the disk and take up real space. cp can handle such files, read the man page about sparse files. Another option i have used in the past it to create a cpio archive using --sparse, copy the cpio archive over and then unpack it, again with --sparse. Andrew