2

I'd like to zerofill all free space in partitions inside a FreeBSD virtual machine.

The virtual machine contains a 512K boot freebsd-boot type slice (hope I got the terminology right here), followed by a 2.0G freebsd-swap slice and a 254.0G freebsd-zfs slice.

# gpart show da0
=>       34  536870845  da0  GPT  (256G)
         34          6       - free -  (3.0K)
         40       1024    1  freebsd-boot  (512K)
       1064        984       - free -  (492K)
       2048    4194304    2  freebsd-swap  (2.0G)
    4196352  532672512    3  freebsd-zfs  (254G)
  536868864       2015       - free -  (1.0M)

It's the layout created by the FreeBSD 10.2 installer by default when picking the "root on ZFS" option.

In the past with UFS I'd simply use mount -t ufs to list all UFS labels and simply create a zero-filled file on these mounts until no space was left.

However, with ZFS I am not longer sure. Now I get:

# mount -t zfs
zroot/ROOT/default on / (zfs, local, noatime, nfsv4acls)
zroot/tmp on /tmp (zfs, local, noatime, nosuid, nfsv4acls)
zroot/usr/home on /usr/home (zfs, local, noatime, nfsv4acls)
zroot/usr/ports on /usr/ports (zfs, local, noatime, nosuid, nfsv4acls)
zroot/usr/src on /usr/src (zfs, local, noatime, nfsv4acls)
zroot/var/audit on /var/audit (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/crash on /var/crash (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/log on /var/log (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/mail on /var/mail (zfs, local, nfsv4acls)
zroot/var/tmp on /var/tmp (zfs, local, noatime, nosuid, nfsv4acls)
zroot on /zroot (zfs, local, noatime, nfsv4acls)

which doesn't give me a clue other than the names, which as I understand are merely a convention (bad style to rely on conventions). And repeating the zero-fill operation on each of those datasets seems a bit silly.

Would it be sufficient then to find all ZFS pools (zpool list -pH|cut -f 1) and look for those in the list mount -t zfs gives me? I.e. ignoring datasets from that ZFS pool.

In short, would it be sufficient to fill the free space on the mount points listed by (using Bash, but likely also works with Zsh):

mount -t zfs|awk '$1 ~ /^'$(zpool list -pH|cut -f 1)'$/ {print $3}'

or does the fact that ZFS has those datasets change what part I need to zerofill before compacting the virtual machine's disk on the host side?

Output when listing the ZFS pools and the mounts:

# mount -t zfs
zroot/ROOT/default on / (zfs, local, noatime, nfsv4acls)
zroot/tmp on /tmp (zfs, local, noatime, nosuid, nfsv4acls)
zroot/usr/home on /usr/home (zfs, local, noatime, nfsv4acls)
zroot/usr/ports on /usr/ports (zfs, local, noatime, nosuid, nfsv4acls)
zroot/usr/src on /usr/src (zfs, local, noatime, nfsv4acls)
zroot/var/audit on /var/audit (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/crash on /var/crash (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/log on /var/log (zfs, local, noatime, noexec, nosuid, nfsv4acls)
zroot/var/mail on /var/mail (zfs, local, nfsv4acls)
zroot/var/tmp on /var/tmp (zfs, local, noatime, nosuid, nfsv4acls)
zroot on /zroot (zfs, local, noatime, nfsv4acls)
# zpool list -pH
zroot   270582939648    1668632576      268914307072    -       0%      0       1.00x   ONLINE  -
|improve this question
  • one complication would be that each of those zfs filesystems may have different quotas or reservations - so zero-filling just one of the filesystems with a quota smaller than the total zpool won't zero-fill the entire disk. check with zfs get quota -t filesystem – cas Dec 3 '15 at 20:51
  • @cas: but wouldn't the root dataset (/zroot in my above example) be a good candidate despite quotas in the child datasets? Of course if that has a quote all bets are off. – 0xC0000022L Dec 3 '15 at 22:18
  • 1
    yep, should be ok unless it has a quota set. – cas Dec 3 '15 at 22:26
  • 2
    Note that creating files with all-zeroes content won't necessarily write all zeroes to disk, because ZFS supports transparent compression. Check zfs get compression zroot -r. If it isn't off everywhere, then writing huge sequences of zeroes will actually write something else to storage. – a CVn Jul 21 '16 at 6:25
  • It seems like so long as compression is off on any file system without a quota set that filling the available space with an all-zeros file should write all zeros all of the empty space. – airhuff Jan 17 '17 at 6:53
1

You're thinking too complicated.

You just need to zero-fill each pool, not each filesystem! (The latter doesn't make a lot of sense in ZFS.)

So just iterate over all pools (via zpool list). For each pool, do the following:

  • Create a new ZFS filesystem with disabled(!) compression
  • Create a new file in that system filled with zeros
  • Sync
  • Destroy that ZFS filesystem

Note that the above algorithm works correctly even in the special case that you have some pool which doesn't contain any filesystem. (either no filesystem yet, or no filesystem anymore)

|improve this answer
0

zpool list will give you a list of the ZFS pools, and zfs list will give you a list of all the ZFS datasets.

|improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.