unimpressive btrfs benchmark results

[Bill Davidsen]

Rex Dieter wrote:
> Neal Becker wrote:
>
>>
> http://www.phoronix.com/scan.php?page=article&item=linux_311_filesystems&num=1
>>
>> I installed using btrfs on my SSD.  Maybe I should be feeling some buyer's
>> remorse?
>
> While being fast is nice, after reading,
> http://btrfs.wiki.kernel.org/
> http://en.wikipedia.org/wiki/Btrfs
>
> It seems clear to me btrfs primary focus is not pure performance, but
> advanced features.  Some choice quotes:
> "... focusing on fault tolerance, repair and easy administration."
> "... intended to address the lack of pooling, snapshots, checksums and
> integral multi-device spanning"
> "it offers improvements in scalability, reliability, and ease of management"
>
> In short, if you expected btrfs to be significantly faster (than ext4, for
> example) for general use, you probably ought to adjust your expectations.
>
If you expect more recent kernels to be faster than previous releases, adjust 
your expectations again. I have noticed this, and it would be nice if Linux 
would bang on people to make things at least as fast in new releases.

I question if journaling is the right thing to do with SSD, putting a beating on 
the device with repeated writes. Is order optimal, given that either directory 
and metadata get written ahead of data, or vice versa and have other issues? 
Have to read up on F2FS and see if that's really a good way to go overall,

In any case and thought provoking data set, suggests areas to test, although the 
number of possible combinations is already large for a small shop.

-- 
Bill Davidsen <davidsen at tmr.com>
   "We have more to fear from the bungling of the incompetent than from
the machinations of the wicked."  - from Slashdot