I think in the case of swap and preserving the no-partitioning thing, ZFS would 
need support either of these:

- The easier of the two, the ability to disable COW on ZVOLs, for performance 
reasons i.e. with swap. Also, if you are willing to trade off reliability for 
performance.
- The harder of the two, the ability to punch a hole into the zpool, which can 
be used for swap like a block device, bypassing most of ZFS' code (apart from 
the RAID-Z/mirror block allocator code, I'd guess).

While for a home user, it wouldn't make much difference if one disk (alone or 
in an unredundant JBOD (aka RAID-0)) is sliced to supply swap, but if you're 
introducing RAID-Z or mirrors, you'll lose the size of swap multiplied by the 
amount of disks affected. For instance, two gigabytes swap in a four disk 
RAID-Z pool results in six gigabytes lost. Hole punching would claim 512 
megabytes on each device and the RAID-Z allocator would spread the blocks 
accordingly, but not doing parity.

Then again, eight gigabytes on a fairly huge RAID-Z is peanuts, but the write 
caches would still be disabled with slicing.

Just an idea, anyway.
-mg

> Swap on ZFS is possible, but I myself would say, a
> plain slice on the
> disk without a ZFS-Block-Device should be faster,
> since ZFS/ZFS-Blockdevice
> is no extra layer in this case. We should ask the ZFS
> peoples about the
> difference in performance between a plain disk slice
> and a ZFS Block-Device.
--

This message posted from opensolaris.org

_______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss

Reply via email to