FreeBSD ZFS vs. TrueOS ZoF benchmarks

[Matthew Macy]

Meaning that it's pre release and the defaults are set so that users can
constructively report bugs.

On Fri, Mar 22, 2019 at 04:50 Paul Pathiakis <pathiaki2 at yahoo.com> wrote:

> Matt,
>
> Meaning?  How does -O0 optimization and INVARIANTS affect this?
>
> Personally, I find everything on Phoronix "out-of-the-box" FreeBSD and
> optimized Linux.  *shrug*  Apples? Meet Oranges.  I make my money as a
> contractor supporting RH/CentOS, but it's always funny to give people the
> heads up on who uses FreeBSD as their starting point and to let them know
> the home network runs on FreeBSD.
>
> When people think Linux (which they believe is the only OSS "Operating
> System" out there)  I have to explain kernel and user space and then
> explain FreeBSD is both and then show them the numerous CVE exploits for
> that year.... 150+ versus ~15.  Faster is not always better, especially
> when you're circumventing standards to get that speed. (I remember the IIS
> vs Apache wars....  Turned out that IIS was not doing things properly and
> circumvented a lot of exploit protections for that speed.)
>
> Building an OS that does everything well OOB, FreeBSD can do that.
> Optimize for application specific....  It usually wins, places or shows.
>
> Sadly, I didn't realize that FreeNAS was using OpenZFS vs the FreeBSD
> ZFS.  Here's my question.... Why?  It was my understanding that SUN made it
> OSS and there are conflicts with the CDDL and GPL.  It seems silly to lose
> performance for no reason.
>
> As for phoronix, I read it for a laugh.  It's funny how so many "Linux is
> everything/rules" people I meet who just use it as a shield and have never
> evaluated the kernels of both and the surrounding userland.  The FreeBSD
> project is tight, goes through a proper QA and release cycle and out pops,
> even a x.0 release, a fully useful new OS version with everything neat,
> tidy, functional and fast.  (So, if FreeBSD can do this, why are all the
> crazies that are producing software screaming AGILE and quick releases
> which still has not solved the problem of crap code?)
>
> Ooops, bit of a rant.... sorry all,
>
> Paul
>
>
>
> On Thursday, March 21, 2019, 12:37:23 PM EDT, Matthew Macy <
> mmacy at freebsd.org> wrote:
>
>
> These were run with ZoF compiled with -O0 and INVARIANTS. Take what you
> read with a grain of salt.
>
> On Thu, Mar 21, 2019 at 09:28 Miroslav Lachman <000.fbsd at quip.cz> wrote:
>
> > There is a benchmark comparing ZFS in FreeBSD 12 with ZFS in TrueOS
> > based on ZFSonFreeBSD 9https://zfsonfreebsd.github.io/ZoF/0
> >
> > FreeBSD ZFS vs. TrueOS ZoF vs. DragonFlyBSD HAMMER2 vs. ZFS On Linux
> > Benchmarks
> > https://www.phoronix.com/scan.php?page=article&item=bsd-initial-zof#=1
> <https://www.phoronix.com/scan.php?page=article&item=bsd-initial-zof&num=1>
> >
> > I am interested if there will be enough testing before replacing the
> > official FreeBSD code base with ZoF. ZFS in FreeBSD 12 is much faster so
> > I am afraid if FreeBSD based on ZoF will be as fast as our current
> > implementation of ZFS.
> >
> > Kind regards
> > Miroslav Lachman
> > _______________________________________________
> > freebsd-fs at freebsd.org mailing list
> > https://lists.freebsd.org/mailman/listinfo/freebsd-fs
> > To unsubscribe, send any mail to "freebsd-fs-unsubscribe at freebsd.org"
>
> >
> _______________________________________________
> freebsd-performance at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-performance
> To unsubscribe, send any mail to "
> freebsd-performance-unsubscribe at freebsd.org"
>