[OpenIndiana-discuss] All of a sudden I'm feeling stupid.... Where does the COW and the newly written data go for snapshots?

[David Brodbeck]

On Wed, May 13, 2015 at 4:33 AM, Udo Grabowski (IMK) <udo.grabowski at kit.edu>
wrote:

> On 13/05/2015 09:20, Hans J. Albertsson wrote:
>
>> This is what it looks like on the FreeNAS box
>>
>> NAME                             AVAIL   USED  USEDSNAP  USEDDS
>> USEDREFRESERV  USEDCHILD
>> Tank/cifs                        8.93T  1.45T     9.18M
>> 1.45T              0          0
>> Tank/cifs at transfer_prep              -  8.24M         - -
>> -          -
>> Tank/cifs at transfer1                  -   966K         - -
>> -          -
>> Tank/cifs at now                        -      0         - -
>> -          -
>> Tank/cifs at auto-20150513.0100-2w      -      0         - -
>> -          -
>>
>> I'd like to see what data is to blame for the 966K reported for
>> transfer1 as "USED"
>>
>
> These are usually access time updates, since the snapshot preserves
> access times. We have a few Megs each day on heavily accessed data.


If you can mount the filesystem 'noatime', that would clear it up pretty
quickly.  I generally do this as a matter of course, these days -- very few
things need atimes and the penalty of adding a write to every read isn't
worth it.  The only things I can think of offhand that might need atime are
mail spools and NNTP.

-- 
D. Brodbeck
System Administrator, Linguistics
University of Washington
GPG key fingerprint: 0DB7 4B50 8910 DBC5 B510 79C4 3970 2BC3 2078 D875