Re: [zfs-discuss] ZVOL vs regular fs — huge difference in data consumption

Etienne Dechamps etienne at edechamps.fr
Fri May 17 12:06:42 EDT 2013


On 05/17/2013 04:59 PM, Daniel Smedegaard Buus wrote:
> But doesn't that mean that on my pool, a file which is 128.3 kB would
> occupy two 128k blocks? I.e. a lot of waste? Or is it smart enough to
> fix that, too?

You're right that in this contrived edge case it would occupy two 128K 
blocks, which is inefficient. There's no such thing as free lunch :) 
(Unless I'm mistaken and it waits for the file to grow larger before 
changing the record size - it's been quite long since I studied this code)

-- 
Etienne Dechamps



More information about the zfs-discuss mailing list