Is ZFS ready for prime time?

krad kraduk at gmail.com
Mon Nov 15 20:33:09 UTC 2010


On 15 November 2010 20:10, Devin Teske <dteske at vicor.com> wrote:

> On Mon, 2010-11-15 at 20:33 +0100, Wojciech Puchar wrote:
> > >>
> > >> _______________________________________________
> > >> freebsd-questions at freebsd.org mailing list
> > >> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> > >> To unsubscribe, send any mail to "
> > >> freebsd-questions-unsubscribe at freebsd.org"
> > >>
> > >
> > >
> > > please elaborate
> > look at archives. i really don't want to repeat the same many times.
> > And anyone that actually have clue about what is computer, disk drive,
> > reliability and algorithms and can think - after reading how ZFS is
> > designed will understand that.
> >
>
> Sounds like FUD.
>
> Like the OP, I too am interested in the current state of ZFS. We've been
> following the threads as far as build 28, and I do indeed see positive
> improvement and continued development. However, is anyone that is
> actively involved in the project able to provide a snapshot opinion of
> the production readiness of ZFS for enterprise deployment?
>
> If the opinion is that ZFS is not ready for production, are there any
> technical explanations as to the efficacy or lack thereof rather than
> the above philosophical FUD which implores the OP to pour over massive
> archives (which can paint an inverse picture because the archives are
> usually filled with a higher number of issues than success stories --
> which is true of nearly ANY mailing-list)?
> --
> Cheers,
> Devin Teske
>
> -> CONTACT INFORMATION <-
> Business Solutions Consultant II
> FIS - fisglobal.com
> 510-735-5650 Mobile
> 510-621-2038 Office
> 510-621-2020 Office Fax
> 909-477-4578 Home/Fax
> devin.teske at fisglobal.com
>
> -> LEGAL DISCLAIMER <-
> This message  contains confidential  and proprietary  information
> of the sender,  and is intended only for the person(s) to whom it
> is addressed. Any use, distribution, copying or disclosure by any
> other person  is strictly prohibited.  If you have  received this
> message in error,  please notify  the e-mail sender  immediately,
> and delete the original message without making a copy.
>
> -> FUN STUFF <-
> -----BEGIN GEEK CODE BLOCK-----
> Version 3.1
> GAT/CS d(+) s: a- C++(++++) UB++++$ P++(++++) L++(++++) !E--- W++ N? o? K-
> w O
> M+ V- PS+ PE Y+ PGP- t(+) 5? X+(++) R>++ tv(+) b+(++) DI+(++) D(+) G+>++
> e>+ h
> r>++ y+
> ------END GEEK CODE BLOCK------
> http://www.geekcode.com/
>
> -> END TRANSMISSION <-
>
>

My gut feeling is no. I wouldn't put it on mission critical stuff yet. Its
not that I have had any major bad experiences (x fingers) but im not aware
of any major deployments of it in the wild. As a result I wouldnt feel safe
being the 1st 8)

What I would advise is to think carefully about what you actually need. If
you dont really need zfs features, then fine go with ufs, as you can always
migrate in the future. However if the features are useful to you and of
enough of a benefit to justify I would advise going for a Solaris platform
of some kind.

If you are doing it on a budget, go for openindiana, but if you have a some
budget, go for the safe option of solaris 10 u9.

It all really depends on your particular circumstances.

A few tips

put the os on its own zpool and the data on its own. That way its fairly
easy to migrate to another os in the future.

go big on ram, the more the better

go for more cores rather than core speed, as zfs will benefit a lot from
more threads (it was designed to run on coolthread archs which have 128+
virt cpus). Also enable hyperthreading  if on intel

look at using ssd for l2arc if you need performance, or are using dedup


More information about the freebsd-questions mailing list