All I can tell you is that the community is dedicated to ensure those 
kinds of limitations don't occur. They work hard to make sure all 
distros and tools align to each other. So far the features that have 
been added or aren't in different distros, can't/don't ever make the 
pools and datasets unreadable. There are things like hot spares or 
sharenfs that are and have been different for awhile. But neither of 
those make the pools unreadable. I think it helps too that the core-team 
for open-zfs is comprised of experts from different distros and 
products. features can only be added to the core. There are also test 
suites available to ensure compatibility.

None of this applies though to those who have gone rogue from the core. 
There are a number of commercial products which have been built from 
openzfs core which have their own features and code sets, which I am not 
sure if they will import in.. But I have seen people take nexenta and 
import into freenas or freenas and import into debian zfsonlinux. I just 
had a partner take really old openindiana and import into syneto. Most 
of them just work. If you try to do this with version 28, it will 
typically want to upgrade your pool to v5000 prior to the import....

 From a community perspective they want it to work.. commercial 
products, don't know. I think the commercial products should have to 
tell you whether or not they are compatible, but a lot of don't even use 
the word "zfs". You usually can tell the products that are dedicated to 
keeping things clean by being listed on the open-zfs.org site.

Some of this is described here

http://open-zfs.org/wiki/Developer_resources

linda



On 4/20/15 1:10 PM, T L wrote:
>
> So if I use a feature which has a new compression scheme (just to pick 
> an example) and then I later move the pool to an OS which doesn't 
> support that scheme, how could this not disable access to the data?
>
> It's that kind of concern that has kept me on 28. I get that all 
> non-Oracle implementations use v5000 but I don't see how, without 
> feature parity being a constraint, one can move amongst open source 
> implementations of ZFS. What am I missing?
>
> Thanks
> Thomas
>
> On Apr 20, 2015 8:22 AM, "Linda Kateley" <lkateley at kateley.com 
> <mailto:lkateley at kateley.com>> wrote:
>
>     Everything beyond v28 is now at 5000. At 5000, feature flags were
>     enabled so that you wouldn't have to worry about features in the
>     different versions. Those features are held in the pool, so if you
>     import into another distro, ZFS will know what features are there,
>     rather than version numbers. Features have to be written to not
>     disable pool operations if that features isn't enabled on the
>     importing distro.
>
>     Once you go up in versions, you can go backwards. The oracle
>     version, last time I looked, was at about 32. So once you import
>     into oracle zfs you can never go back again. But you can import a
>     free and open pool into oracle zfs.
>
>     linda
>
>     On 4/19/15 8:28 PM, T L wrote:
>>
>>     I get those messages on v28 pools, made with ZFS for Linux or
>>     FreeBSD or older versions of FreeNAS.
>>
>>     I'm hesitant to move beyond v28, even though I'd like some of the
>>     newer features, because I am not clear on the compatibility
>>     matrix among non-Oracle ZFS implementations. Version 28 seems to
>>     be the last version that all extant ZFS implementations can read.
>>     I'd be very willing to give up Oracle compatibility but don't
>>     want to lock myself out of moving between ZFS on Linux and *BSD.
>>     If anyone else has thoughts or experience to share here, that'd
>>     be mist welcome!
>>
>>     Thomas
>>
>>     On Apr 19, 2015 6:37 PM, "Clug" <tclug at freakzilla.com
>>     <mailto:tclug at freakzilla.com>> wrote:
>>
>>         Oh, in case anyone was curious, this is what zpool upgrade
>>         returned, so these features for some reason weren't enabled
>>         before.
>>
>>             clug at rooster:/home/clug> sudo zpool upgrade
>>             This system supports ZFS pool feature flags.
>>
>>             All pools are formatted using feature flags.
>>
>>
>>             Some supported features are not enabled on the following
>>         pools. Once a
>>             feature is enabled the pool may become incompatible with
>>         software
>>             that does not support the feature. See zpool-features(5)
>>         for details.
>>
>>             POOL  FEATURE
>>             ---------------
>>             media
>>                   spacemap_histogram
>>                   enabled_txg
>>                   hole_birth
>>                   extensible_dataset
>>                   embedded_data
>>                   bookmarks
>>
>>         --
>>         _______________________________________________
>>         TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
>>         tclug-list at mn-linux.org <mailto:tclug-list at mn-linux.org>
>>         http://mailman.mn-linux.org/mailman/listinfo/tclug-list
>>
>>
>>
>>     _______________________________________________
>>     TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
>>     tclug-list at mn-linux.org  <mailto:tclug-list at mn-linux.org>
>>     http://mailman.mn-linux.org/mailman/listinfo/tclug-list
>
>
>     _______________________________________________
>     TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
>     tclug-list at mn-linux.org <mailto:tclug-list at mn-linux.org>
>     http://mailman.mn-linux.org/mailman/listinfo/tclug-list
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mn-linux.org/pipermail/tclug-list/attachments/20150420/8dc43e85/attachment-0001.html>