error writing 'wipefile': No space left on device

I have a question. I have Vmware workstation and one of my Vms is nextcloud with ubuntu 1804.

I like to have a snappy instance so I use these commando's found here:

sudo e4defrag /
dd if=/dev/zero of=wipefile bs=1M; sync; rm wipefile
sudo vmware-toolbox-cmd disk shrinkonly

For a while this worked fine. But now on step 2: dd if=/dev/zero of=wipefile bs=1M; sync; rm wipefileI get error:

64172396544 bytes (64 GB, 60 GiB) copied, 56.5438 s, 1.1 GB/s
ncadmin@cloud:~$ error writing 'wipefile': No space left on device

But when I do a: df -Th I have plenty of space:

Filesystem Type Size Used Avail Use% Mounted on
udev devtmpfs 3.9G 0 3.9G 0% /dev
tmpfs tmpfs 796M 1.4M 794M 1% /run
/dev/mapper/ubuntu--vg-ubuntu--lv ext4 103G 39G 60G 40% /
tmpfs tmpfs 3.9G 0 3.9G 0% /dev/shm
tmpfs tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs tmpfs 3.9G 0 3.9G 0% /sys/fs/cgroup
/dev/sda2 ext4 976M 157M 753M 18% /boot
/dev/loop0 squashfs 89M 89M 0 100% /snap/core/739 6
/dev/loop1 squashfs 55M 55M 0 100% /snap/lxd/1210 0
/dev/loop2 squashfs 55M 55M 0 100% /snap/lxd/1198 5
/dev/loop3 squashfs 90M 90M 0 100% /snap/core/771 3
tmpfs tmpfs 796M 0 796M 0% /run/user/1000

I only have basic linux understanding

1 Answer

No space left on device was expected. This command

dd if=/dev/zero of=wipefile bs=1M

reads from /dev/zero and writes to wipefile. The special file /dev/zero cannot be depleted. The command runs until wipefile cannot grow anymore (or until you manually terminate it first).

So you clogged the filesystem, there was indeed no space left on device. Then you flushed cache with sync and removed the file with rm. The file was exactly 64172396544 bytes before removal, this space was reclaimed when you removed the file. Now you have about 60 GiB available on /. All fits.

The point of creating this file was to fill available space with zero bytes; the point of syncing was to make sure all zeros actually got to the device; the point of removing the file was to formally mark zeroed blocks as empty again. You didn't need the file itself; you needed to write zeros to "empty" blocks. A file that later gets removed is just a simple way to do it.

I guess vmware-toolbox-cmd can now do its job: shrink. I don't know this tool at all but the idea of filling empty space with zeros is quite sane when you keep (or want to keep) the entire filesystem as a file and you want the file to be as small as possible. The main filesystem of your VM is just a file from the point of view of the host, right? Compare this answer of mine where the same method is used to prepare a filesystem to be effectively stored as a regular file. My code explicitly says

## Long wait here. Expect "no space left on device".