I want to minimize the amount of space my virtual disk is using, by zero'ing out the data on any deleted files. The virtual disk is a VDMK running ubuntu with a single Ext4 partition on VirtualBox.
What is the best way to find any deleted files, and zero them out, so that when I export the appliance the disk size is only existing files on the disk?
If you are using a filesystem that supports it then you can use zerofree.
If you are not then you basically need to just use dd to write zero to a file on each partition until there is no space left. Then delete the file you filled with zeros.
For each file system, do like
sudo dd if=/dev/zero of=/zero
# ...wait until the filesystem is full...
sudo rm /zero
That will create a file over all free space on the filesystem, filled with zeros. When you delete it, there will be only zeros left on partition. However, I guess you need to gzip the virtual disk anyway, as zeros take up space anyway. I am not sure how smart virtualbox disk management is.
Dump and restore would work.
If you haven't deleted the files yet, you could do something like this:
shred --iterations=0 --exact --zero --remove file
For this to work, you have to use thin provisioning. Or am i wrong there?
