ZFS and disk usage

[Ronald Klop]

On Fri, 13 Apr 2012 17:27:10 +0200, Johannes Totz <johannes at jo-t.de> wrote:

> On 13/04/2012 14:28, Mark Schouten wrote:
>> Hi,
>>
>> Op Vrijdag, 13-04-2012 om 15:11 schreef Volodymyr Kostyrko:
>>>> These are fiesystems that are created with the following
>>>> command. zfs create -V ${size}GB ${ZFS_ROOT}/${diskname}
>>>
>>> `zfs create -V` withous `-s` creates reserved volume that eats all
>>>  needed space immediately. Technically zfs pool is filled only for
>>> 23%, but logically you have only 138G left unassigned.
>>
>> I understand. However, the created volumes should use a total of
>> 1211GB. That's not 1.6TB like zfs list says. But 1211 + 431
>> (referred) does come close to 1.6TB.n And 1.6 TB still isn't the
>> 1.77TB that's in the zpool.
>>
>> I have this feeling that zfs has reserved the space for each volume,
>> but counts data written to the volumes in usage of the main
>> filesystem. Mainly because zfs list shows me that the volumes have
>> only 16KB referenced, where /storage has 431GB referenced.
>
> Without checking the numbers myself...
> Note that zpool and zfs do not agree on (free) space accounting: zpool
> shows "raw" space, whereas zfs includes metadata overhead for itself.
>
> Small rant: I dont understand why zpool and zfs show different things.
> If you have an integrated storage stack then why not show consistent
> numbers? Is there any use for this extra (mis-)information that
> zpool-vs-zfs provides?
>

It isn't integrated that far as you might think. That is why you have  
zpool and zfs tools. These are 2 separate things. You can also put a zvol  
+ ufs or zvol + iscsi on the zpool and not use zfs. How would zpool know  
about ufs or iscsi usage? Or how would ufs know about the underlying  
zpool+zvol?

Ronald.