Slashdot Mirror

OpenZFS and Oracle ZFS have diverged a bit. The on-disk pool contains a version number which identifies with certainty whether you can import it on a given implementation, so there's at least no chance of mistaken mis-importing & data loss from that. They're interoperable for pools that aren't upgraded past the highest pool version supported in the final CDDL release of Oracle ZFS. Beyond that, they won't work.

Oracle ZFS has since added file-level encryption. The encryption and the on-disk structure aren't readable by OpenZFS. OpenZFS has incremented the pool version number by a large jump (5000) past the last Oracle ZFS version and has fixed & enhanced some things in such a way that the on-disk isn't compatible with Oracle ZFS. For info about OpenZFS version & feature flags, see http://open-zfs.org/wiki/Featu...

I don't think it would take a tremendous amount of effort to merge the functionality one way or the other if the licensing issues were solved, but they're definitely not on-disk compatible if you're running the latest pool version supported by either release.

ZFS enhancements come to mind.

After the split with OpenSolaris, ZFS development moved to the Illumos project and OpenZFS grew out of it.

ZFS on Linux (http://zfsonlinux.org/) is a great option via the great work done by Lawrence Livermore National Laboratory Also, have a look at: http://open-zfs.org/wiki/Distr... and http://open-zfs.org/wiki/Compa... for solutions where ZFS is integrated into various solutions.

ZFS on Linux (http://zfsonlinux.org/) is a great option via the great work done by Lawrence Livermore National Laboratory Also, have a look at: http://open-zfs.org/wiki/Distr... and http://open-zfs.org/wiki/Compa... for solutions where ZFS is integrated into various solutions.

> Because ZFS on linux is not yet ready for serious use

--That is not my experience. Killer-app for Linux is ZFS+Samba, and I have had that going with decent performance (for home / small business use, at least) for the last 2-3 years.

--This is with an older quad-core 2.4GHz COTS PC box with 6GB DDR2 RAM running Xubuntu 14.04--64--LTS and standard SATA-2 hard drives connected to (2) inexpensive 4-port PCIe cards. With standard 1500-byte Ethernet frames, I can almost saturate the link (~100-120MB/sec) running an FTP transfer from a RAIDz10 with noatime (2x2 disk pool.)

--The limiting issue is usually the write speed on the receiving side, unless it's going to SSD. (Writes to the pool are admittedly a bit slower IIRC, but zfs *is* doing auto-checking for every write.) But in general, **you don't run ZFS for max speed**, you run it for Reliability -- and features like "copies=2" + snapshots + filesystem-level compression. As well as the occasional quick RAID rebuild.

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

--What some folks might want to do to speed things up is some research; I have some custom stuff going on in /etc/rc.local (sysctls to speedup I/O, blockdev --setra 8192 /dev/sd* , Gig ethernet speedup script, rmmod unused modules, stop all unneeded services, etc. Optimizations are available upon request.) Bog-standard Linux install is not optimized for speed, usually. Plus if you're running all your pool drives off motherboard SATA ports, it can limit your speed. On my rig, the "mirror" drives in the pool are connected to the 2nd SATA card.

--Also - if you want best speed from ZFS, most advice is to run mirrored pools - *not* RAIDZ. I switched out my original 500GBx6 RAIDZ2 to equivalent RAIDz10 (with 2x2 newer 2TB WD Red drives, which aren't even considered hi-speed drives) and the pool performance improved. If I expand the pool to 6 drives, it should pick up even a little more speed. If I wanted to spend more money, I could put a SAS card in and run that instead of SATA.

--An interesting feature of ZFS, BTW - I started out with 2x2TB drives, unmirrored, making a non-RAID writable space of ~4TB. After saving up for a couple of months and buying 2 more drives, I was able to add mirrors to both disks in the pool ON-THE-FLY and converted the pool in-situ to RAIDz10. I did a "burn-in" test (R/W all sectors) on the new drives 1st, and I consider WD to be pretty reliable to begin with, but still - I don't know any other filesystem that you could do that with.

zfs sendstreams contain serializations of DMU objects, which are the fundamental objects in zfs; in sendstreams, DMU objects are not compressed, and they are not individually checksummed.

You can verify that with zstreamdump if you like, or read more details here http://open-zfs.org/wiki/Docum...

There are forthcoming changes to the zfs sendstream which will compress and add checksums within the stream itself, but that's to protect the integrity of the *stream* rather than of use to the receiver, which simply adds DMU objects to a dataset or zvol according to the policies of the receive side. More specifically, zio options (such as compression, copies, deduplication, and so forth) are presently only ever even knowable by the receiver when properties are in the send stream, which you only get with "zfs send -p" or "zfs send -R" (and the receiver can ignore them in favour of its own properties), and AFAIK there is no plan whatsoever to change that.

Nope. Not 'owned'. It's covered under the CDDL and developed by a group that isn't associated with Sun. Open-ZFS.

They would also be sued pretty quickly by Oracle. Clearly not an option.

Your conclusion is a bit hasty and unwarranted. I am not going to tell you that Oracle CANNOT sue anyone for any trumped-up reason, but ZFS is licensed under the Common Development and Distribution License (CDDL) and is open source. For linux, there is an issue with how CDDL plays with GPL, so no distro has yet bundled ZFS with linux. Linux users, however, can themselves pick up "ZFS on Linux" and install it themselves without violating either the CDDL or GPL.

But OSX is not GPL. Other systems that are not GPL bundle ZFS, and are not sued. For example, FreeBSD comes with ZFS, and there are a number of other systems, such as FreeNAS, PS-BSD, illumos and nexenta.

See OpenZFS.

Check out: http://open-zfs.org/wiki/Main_...
http://dtrace.org/blogs/ahl/20...

ZoL is very active and very up-to-date. All the versions and compatibility is in sync with Illumos (the main source of OpenZFS) and FreeBSD. You can create and move zpools between these 3 platforms seamlessly.

2 of the main founders and creators of ZFS itself (who used to work for Oracle and wrote ZFS) who now work for Delphix and continue to improve OpenZFS (started with the last open release of Oracle ZFS) in Illumos and have actually made it better than the now closed Oracle ZFS.

See how OpenZFS is actually better than Oracle ZFS now:
http://dtrace.org/blogs/ahl/20...

Actively adding new features: https://github.com/zfsonlinux/... (Largeblock support to match the newer Oracle ZFS)

Etc...

Maybe if we pray to the ZFS gods Matthew will start implementing block pointer rewrite some day!

Anyways, OpenZFS is very active and kicking, and that includes the Linux port.

hope that this nice umbrella will evolve into a single point of access

If you find time, I recommend listening to bsdtalk227 (listed under Publications ).

https://twitter.com/grahamperrin/status/380395699734466560 quoting plus hashtag from a Delphix blog: " To some degree, #OpenZFS is just putting a name to what we have already been doing as a community ".

Sometimes these things turn into just another layer of non-information.

I understand your concern.

The first of the three goals of OpenZFS is to raise awareness of the quality, utility, and availability of open source implementations of ZFS. As an end user, very much into awareness-raising of ZFS, I'll occasionally edit (and/or discuss in IRC) wherever I feel that the value of something thats in the wiki is not immediately clear. But I'm neither a developer nor a typical end user, so there'll be large areas that are beyond me. Maximising the value of contributions to the wiki should be very much a collaborative effort...

hope that this nice umbrella will evolve into a single point of access

If you find time, I recommend listening to bsdtalk227 (listed under Publications ).

https://twitter.com/grahamperrin/status/380395699734466560 quoting plus hashtag from a Delphix blog: " To some degree, #OpenZFS is just putting a name to what we have already been doing as a community ".

Sometimes these things turn into just another layer of non-information.

I understand your concern.

The first of the three goals of OpenZFS is to raise awareness of the quality, utility, and availability of open source implementations of ZFS. As an end user, very much into awareness-raising of ZFS, I'll occasionally edit (and/or discuss in IRC) wherever I feel that the value of something thats in the wiki is not immediately clear. But I'm neither a developer nor a typical end user, so there'll be large areas that are beyond me. Maximising the value of contributions to the wiki should be very much a collaborative effort...

https://github.com/zfsrogue/zfs-crypto

In the ZFSonLinux area at https://github.com/zfsonlinux/zfs/issues/494#issuecomment-23652335 it's noted that the zfsrogue code is encumbered and so, will not be used.

There's an earlier comment https://github.com/zfsonlinux/zfs/issues/494#issuecomment-7158618 and a corresponding note in the OpenZFS wiki: The early ZFS encryption code published in the zfs-crypto repository of OpenSolaris.org could be a starting point

http://open-zfs.org/wiki/Developer_resources#Implementation_documentation notes that the ZFS On-Disk Format document is "a good overview, but sorely outdated". Of possible interest: Max Bruning's weblog: ZFS Raidz Data Walk (2009)

FAQ much? There is no central source repository for OpenZFS. Each supported operating system has it's own repository. The previous also has a link to the source tree for each of the supported projects under the umbrella.