[zfs-discuss] Re: Power-Loss protection on SSDs and sync=disabled

devsk devsku at gmail.com
Wed Jan 1 14:12:43 EST 2014

On Wednesday, January 1, 2014 8:16:47 AM UTC-8, Matthew Thode 
(prometheanfire) wrote:
> On 12/31/2013 03:20 PM, devsk wrote: 
> > I thought only M500 had the power loss protection (super caps). May be I 
> am 
> > wrong. 
> > 
> > The analysis mentioned only increasing number of CRC errors with power 
> > cycles and nothing about actual corruptions on M4. How superficial? CRC 
> > errors do not mean actual corruption of the user data. 
> > 
> > -devsk 
> >
> the s3700 has them as well that I know of. 
> This may be good reading as well. 
> http://lkcl.net/reports/ssd_analysis.html 
I thought I was commenting on that "analysis"...:-)

Its a bogus study designed to prove a well-set hypothesis from their 
experience that Intel drives are good, which may very well be a valid 
conclusion but the analysis is incomplete for "buy only Intel drives" 
conclusion. e.g.

1. Big fail for me was that Crucial m4 is not marketed with "power loss 
protection". m500 is. Lets test that and see.
2. They never mention any real corruptions with m4, only that CRC errors 
kept increasing. Why did they single out m4 to base the analysis off of 
SMART attributes only instead of basing it on actual data corruptions? Let 
me guess...because this was the only drive which probably gave Intel a run 
for its money and that went against their basic premise. May be if the m4 
firmware did not report the CRC errors, they would be happy?
3. The analysis did not include so many other good drives with power loss 
protection in the market available when the test was carried out.


To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+unsubscribe at zfsonlinux.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.zfsonlinux.org/pipermail/zfs-discuss/attachments/20140101/2496e08d/attachment.html>

More information about the zfs-discuss mailing list