ZFS on Core Storage with Fusion drive

[Alex Blewitt]

http://jollyjinx.tumblr.com/

Looks like this provides some optimisations that zfs' cache device would do, but based on what blocks are read.

Alex

Sent from my iPhone 4S

[Fastmail Jason]

Yeah, I was reading this last night and my brains been whirling all night, will start playing with it tomorrow.


--
Jason Belec
Sent from my iPad

> --
>
>
>

[Graham Perrin]

Patrick's answer at 

<http://apple.stackexchange.com/a/70113/8546>

<http://apple.stackexchange.com/questions/69538/can-i-make-a-diy-fusion-drive?lq=1#comment81888_69538> comment from me. 

There's an assumption that routine scrub may negate some of benefits of auto-tiering. Discussion at/around 

<http://zevo.getgreenbytes.com/forum/viewtopic.php?p=1413#p1413>

[Daniel Becker]

ZFS is a copy-on-write file system; that means that on-disk data is never updated in-place: every time a block of data is modified, it is effectively copied to a new location. So for writes, this will inherently negate any auto-tiering. 

Furthermore, I seem to remember reading somewhere (Ars?) that the HFS+ layer uses CoreStorage-specific IOCTLs to actually initiate the tiering; if that is indeed the case, ZFS will never generate these, and the net effect will be that you have ZFS running on what is effectively a concatenated device with very non-uniform performance characteristics (depending on whether any given block address is located on the SSD or the HDD).

--