Linus Torvalds Doesn't Recommend Using ZFS On Linux

[blackiwid]

Lots of people use FreeBSD. So much so it's on the top cloud providers in the world such as Amazon, Azure, Google Cloud and others.

Of course when I speak about Users I speak about their own machines / Desktop, I don't go around and say hey "Linux" is on most computers because, it drives the Internet and all 10 quadrilliards of smartphones use it. The desktop is what matters and there Linux has like 2% or something und BSD like 0.1% if not less.

[ryao]

That is wrong I don't get where you get the idea that there are no rulings over GPLv2 and patents.

http://www.informit.com/articles/article.aspx?p=2449028

There are rulings in court like this that GPLv2 contains implied patent license. So GPLv2 does not need a patent grant due to how the license is written. You cannot push patent infringement to breaching your requirements under copyright law that is how GPLv2 implied patent license works so patents laws don't superseded copyright in fact copyright can be written the written in way that superseded all patents laws. Funny enough that was in GPLv2.

This is the first that I have heard of this. Neat.

That said, I doubt that case would mean much to Oracle. They similarly lost the first case against Google because the first court had common sense, but won on appeal. Unless we have a Supreme Court ruling, there is room for Oracle to maneuver if it wants, plus this is just in the US. If they wanted, they could go after anything being distributed worldwide in any jurisdiction. ZFS is not any worse off than Linux is against that threat.

This statement if you like it or not is true. We need Oracle legal department to clearly state if they have traded the copyright away and who are the patent holders over the SUN ZFS code. Maybe Oracle made ZFS closed source again because they don't have the patents or the copyright to legally distribute it under CDDL. The only party who can answer what the state of play is Oracle.

It does not matter who owns the copyright on ZFS because the CDDL is a perpetual license. Once the copyright holder makes a copy of code available under it, that copy (and all copies of that copy) stays under it.

If you have heard what I have heard from both former and current Oracle employees, you would not entertain any of these theories. Making things closed source was the whim of a single executive. The netapp lawsuit ended because netapp had zero interest in a fight with Oracle. It only started because Sun tried to talk to netapp about royalties for the ZFS patents. Oracle had zero interest in that and instead wanted royalties from Java. You know how that went. The executive in question is not a generous man and would never give anything away unless he got something in return. The idea of him giving away the copyright to anything Oracle owns is absurd.

That said, Linus’ concern is about Oracle reinterpreting things as he stated in his second email. He never once brings up this netapp theory. The concern about Oracle reinterpreting can be applied to them reinterpreting things to go after Linux itself. It is rather unlikely that they would bother with ZFS. The big money is with Linux, as Microsoft found. ZFS also has a strong license intended to protect it, plus numerous companies who feel confident in that. Some are listed here:

http://www.open-zfs.org/wiki/Companies

This is missing the implied patent license of GPLv2 has been used against Microsoft since it has been proven to exist. Some Android makers have used this to get out of paying Microsoft any more money.

Whether it applies to patents not relevant to source code contributions or to patents acquired after the fact is an issue with such a theory. Oracle gained a number of UNIX patents for things it never touched in Linux. Anyway, if we are entertaining Linus’ remark about Oracle reinterpreting things, then anything we say about Linux being safe from Oracle is moot given that Oracle can just reinterpret things to suit them in court. The logic fits.

That being said, I think Microsoft joining OIN mattered much more for Android vendors getting out of their agreements to pay royalties to Microsoft than that court ruling:

https://azure.microsoft.com/en-us/blog/microsoft-joins-open-invention-network-to-help-protect-linux-and-open-source/

With the FAT patents for example, there could not be an implied license for it because Microsoft never contributed to that code. Speaking of which, they seem to have voluntarily dropped the ability to assert those against Linux:

https://www.zdnet.com/article/microsoft-readies-exfat-patents-for-linux-and-open-source/

[ryao]

So let's quote the article you link to that describes the crux of the matter:


The majority of the article and your arguments are straw-man arguments since binary distribution and source code availability for the end user is not what's on the table here, the problem is that Linus cannot distribute the Linux Kernel source code under GPLv2 if he includes the ZFS source code to the kernel source code tree.

Linus does not find that to be an issue from what he has told me in person. He already has stuff in his tree under licenses other than the GPL.

By the way, for what it is worth, I hold copyright over part of OpenZFS. You do not have any standing to be telling me about OpenZFS’ copyright situation, especially after the years of arm chair remarks by guys like you led me to speak to several attorneys on this topic. You also just dismissed a legal opinion by the most preeminent legal scholar on OSS as strawmans. It is rather ridiculous.

Whether or not Nvidia, IBM or Broadcom are releasing and code from their binary distribution is a separate issue completely from what Linus have to do by including ZFS.

These days, no one is asking him to include ZFS as far as I know. He already made his position on that clear. The question that caused him to state unrelated things had to do with a symbol export removal that caused a headache for OpenZFS.

That being said, I can say as one of the OpenZFS developers that I am not aware of any of us being interested in getting ZFS into his tree. It would make more work for us. I do not understand this presumption that people seem to make about getting into his tree being the only thing that matters.

[oiaohm]

This is the first that I have heard of this. Neat.

That said, I doubt that case would mean much to Oracle. They similarly lost the first case against Google because the first court had common sense, but won on appeal. Unless we have a Supreme Court ruling, there is room for Oracle to maneuver if it wants, plus this is just in the US.

There are other rulings in Germany and 10 other countries that all rule the same way. Its turns out we had something wrong. When you read through the agree conventions between countries over copyright and patents there is something interesting.

Note how prior art that is a copyrighted documented can over turn a patent. People presumed patents over rules copyright. Reality copyright over rules patents. Due to this being convention arguing over GPLv2 implied patent license to be ultra horrible can take you into the international court over trade that make the Supreme court of the USA look like a minor problem. So a GPLv2 case over the implied patent part of license if ruled against by the USA Supreme Court can go to the international court where there is already a ruling in GPLv2 favour.

Basically you would be a idiot to push a GPLv2 implied patent license case up though the USA court system no matter how it rules you have to go to international court in the end and lose..

If they wanted, they could go after anything being distributed worldwide in any jurisdiction. ZFS is not any worse off than Linux is against that threat.

CDDL has not been tested in multi jurisdictions on patents where GPLv2 has been. GPLv2 implied patent license is tested and proven in the highest court on earth and there is no loophole. Yes USA people forget there is a court above the USA Supreme Court that can at times over rule its rulings.

It does not matter who owns the copyright on ZFS because the CDDL is a perpetual license. Once the copyright holder makes a copy of code available under it, that copy (and all copies of that copy) stays under it.

It matter when you have license conflict and you are working out how you can interpret license based on intent of the license(spirit). Without being able to ignore sections of CDDL or GPLv2 they are not compatible.

Making things closed source was the whim of a single executive.

That does not answer why the whim and if there was any deal either. Lower level staff don't need to be told stuff. Legal department would have had to be told stuff.

The executive in question is not a generous man and would never give anything away unless he got something in return.

That the point the netapp settlement there was the right to use the netapp patents and other tech. So he was getting something in return what did he trade to pay for that.

That said, Linus’ concern is about Oracle reinterpreting things as he stated in his second email.

True and this is because CDDL is vague in places and incompatible.

He never once brings up this netapp theory.

That Linus did not raise it does not mean its not a factor. Remember getting a clean approval from Oracle as Linus requested would kill the possible netapp problem.

The concern about Oracle reinterpreting can be applied to them reinterpreting things to go after Linux itself. It is rather unlikely that they would bother with ZFS.

https://en.wikipedia.org/wiki/Oracle_Corp._v._SAP_AG

Oracle making things up after the fact is well documented.

ZFS also has a strong license intended to protect it, plus numerous companies who feel confident in that. Some are listed here:
http://www.open-zfs.org/wiki/Companies

None of those have a major legal firm who would have the resources to defend ZFS if they had to. Feeling confident and having the resources to back it up is another matter.

Whether it applies to patents not relevant to source code contributions or to patents acquired after the fact is an issue with such a theory. Oracle gained a number of UNIX patents for things it never touched in Linux.

Really would pay to check who Oracle got those patents from. Some of Oracle Unix patents trace back to Novell/SCO who both got their ears boxed in for going after Linux by IBM. IBM was the first to pull out the GPLv2 implied patent license. So Oracle using what they have against Linux and lose due to prior cases around the patents with related legal records would be facing triple damages and costs if they loses. Yes Oracle losing would be basically 100 percent. Linux kernel is very well protected by parties with lots of money if they get serous.

Anyway, if we are entertaining Linus’ remark about Oracle reinterpreting things, then anything we say about Linux being safe from Oracle is moot given that Oracle can just reinterpret things to suit them in court. The logic fits.

With CDDL yes Oracle could reinterpret things. With the Linux kernel mainline GPLv2 prior cases that Oracle legal department would be nuts to go anywhere near it. Some of the reason Oracle is so sure they can push Java LGPL so hard against Google is that Oracle legal department has a very good idea how dangerous GPL licenses are.

Linus logic is right no matter how you argue. GPLv2 position is solidly legally proven. CDDL is not legally proven. Those companies trusting open-zfs are basically doing a leap of faith.

[ryao]

That the point the netapp settlement there was the right to use the netapp patents and other tech. So he was getting something in return what did he trade to pay for that.

Most of Netapp’s customers were using Oracle software and would have dropped Netapp if Oracle started discouraging the use of Netapp with Oracle software. Considering that, Netapp was in an unenviable position and they were willing to give Oracle things to get out of it without getting much in return. Netapp was only in this situation because it wanted to avoid paying money to Sun. It was a jerky maneuver that they did to say no to Sun’s request for patent royalties (which itself was jerky). They could have simply had their lawyers talk more.

As far as I know, Oracle was able to settle the lawsuit without giving very much. They did not want to divert resources from the Java lawsuit and Netapp did not want to be in the lawsuit. Netapp did get what they originally wanted though, which was not having to pay royalties over the ZFS patents.

Yes USA people forget there is a court above the USA Supreme Court that can at times over rule its rulings.

There is no court higher than the US Supreme Court in the US according to the US constitution. It’s decisions are final. If you look at “international courts’” decisions, you would find the decisions are non-binding (and are often ignored by the losing parties). Winning one is just PR for the winner. As far as I know, the US Supreme Court’s rulings can only be overridden by new US Supreme Court rulings or a US constitutional amendment. It is possible for the US Congress to authorize external courts to rule on things (just as it is able to make courts for various things), but such authorization never preempts the US Supreme Court.

That Linus did not raise it does not mean its not a factor. Remember getting a clean approval from Oracle as Linus requested would kill the possible netapp problem.

https://en.wikipedia.org/wiki/Oracle_Corp._v._SAP_AG

Oracle making things up after the fact is well documented.

Linus is concerned about Oracle being a problem, not netapp.

None of those have a major legal firm who would have the resources to defend ZFS if they had to. Feeling confident and having the resources to back it up is another matter.

We have the SFLC and plenty of companies willing to work together if there is an issue. Why do you think we have OpenZFS?

Really would pay to check who Oracle got those patents from. Some of Oracle Unix patents trace back to Novell/SCO who both got their ears boxed in for going after Linux by IBM. IBM was the first to pull out the GPLv2 implied patent license. So Oracle using what they have against Linux and lose due to prior cases around the patents with related legal records would be facing triple damages and costs if they loses. Yes Oracle losing would be basically 100 percent. Linux kernel is very well protected by parties with lots of money if they get serous.

They got many more from Sun. The ZFS patents get plenty of attention, but they are just a fraction of what they got from Sun.

As for Oracle losing, they would lose if they went after ZFS too. That is what my legal counsel believes anyway.

With CDDL yes Oracle could reinterpret things. With the Linux kernel mainline GPLv2 prior cases that Oracle legal department would be nuts to go anywhere near it. Some of the reason Oracle is so sure they can push Java LGPL so hard against Google is that Oracle legal department has a very good idea how dangerous GPL licenses are.

Linus logic is right no matter how you argue. GPLv2 position is solidly legally proven. CDDL is not legally proven. Those companies trusting open-zfs are basically doing a leap of faith.

Linus’ logic is that Oracle can change whatever everyone thought was correct. By that logic, nothing is safe. Honestly, software patents are a really stupid invention that the US Supreme Court gave us (in part thanks to IBM). If anything, I have learned to expect them to mess up any ruling on technology, which they did with Oracle vs Google. If I am willing to let myself accept what Linus is saying, then I might as well not use a computer because everything would be suspect. This is a reductio ad absurdem. My point is that you can live being afraid of everything or just listen to legal counsel and go on with your life. If fear decided how we did things, no one would use Linux because those court cases on the GPL and patents only existed because people had used it thinking that it would be fine.

[k1e0x]

There are no rulings over the GPL and patents. The GPLv2 also has no patent grant, unlike the CDDL. Oracle is also holding plenty of patents from Sun on UNIX in general. If they wanted to use the nuclear option, ZFS is in a much better position than Linux due to the patent grant of the CDDL.

Your remark “without SUN legal information that Oracle holds how many more parties are there like this” is FUD that can be applied to *anything*. Microsoft has a few patents that it has used against the Linux kernel to collect $1 billion in annual royalties from Android phone makers, so it is clear that such FUD could be easily applied to Linux too. Oracle has enough UNIX patents that they likely could do the same or worse with Linux if they tried.

This.

The GPLv3 was written for this exact case. And it isn't just Oracle. IBM (RedHat), AT&T, Novell, Microsoft and others own huge amount of Unix patents.

I'm convinced that the reason Oracle hasn't already gone after commercial implementations of OpenZFS already is that they would lose. They pay people 24/7 to look at issues like this.

If they really wanted to pull the nuclear option one could also challenge the assumption software warranty exceptions are valid.

[F.Ultra]

Linus does not find that to be an issue from what he has told me in person. He already has stuff in his tree under licenses other than the GPL.

And are those other licenses incomaptible with GPLv2? No they are not so what the point of this argument???

By the way, for what it is worth, I hold copyright over part of OpenZFS. You do not have any standing to be telling me about OpenZFS’ copyright situation, especially after the years of arm chair remarks by guys like you led me to speak to several attorneys on this topic. You also just dismissed a legal opinion by the most preeminent legal scholar on OSS as strawmans. It is rather ridiculous.

No I didn't dismiss the legal opinion of the "most preeminent legal scholar on OSS", I quoted him verbatim. The strawman is you bringing up that article in connection with including ZFS in the Linux Kernel source code tree when it's all about the legal possibilities of including ZFS as a binary distribution together with the Linux kernel.

And what does your copyright on parts of OpenZFS matter? We are talking about the license that code is licensed under, and how that license is incompatible with GPLv2. We are not arguing over your contributions or code.

If you one day own the copyright to 100% of the ZFS code then it does matter because then you have the legal right to relicense it, but we are not there now are we.

[allquixotic]

I posted this on Slashdot, but I'll reiterate it here. I have nothing to say about licensing; the situation is what it is. This is a purely technical argument for why ZFS is still unique, and fills a common need in a way that no in-tree FS does.

There are a lot of comments asking "why ZFS" but most of them don't really understand the main killer feature of ZFS (and ZFS on Linux): the ability to efficiently use tiered storage.

See, there's currently a problem overall with the storage industry. Flash storage, aka SSDs, is hideously expensive per gigabyte. Magnetic storage, aka HDDs, is hideously slow in terms of IOPS. For difficult workloads that require an optimization of server purchase price, high IOPS, and large quantities of local storage, Tiered Storage is the only real option.

This allows you to buy both: (1) relatively inexpensive, high write endurance but fairly low capacity SSDs -- usually on the order of 128 to 512 GB, depending on the size of the HDDs behind them; and (2) relatively inexpensive, high capacity but slow HDDs -- usually 8 TB or larger -- and combine them into one logical block device that *behaves* as if it were an SSD with many terabytes of storage. You get about 98% of the IOPS performance of the SSDs, while all the data ultimately persists to the HDDs behind the scenes. This is remarkably good for large databases and file storage servers, and the price of building all of that capacity in datacenter-grade SSDs is going to run you about $1000 more per terabyte of capacity, assuming RAID-1 redundancy.

With ZFS, you can set up your ZFS Intent Log (ZIL) -- basically a write buffer for the HDDs -- on a partition about 25-50% of the SSD capacity (depending on how write-intensive your workload is), and set up the ZIL in RAID-1 mode for data safety. ZFS will then efficiently create large batch sequential writes to the HDD that convert what could be thousands of SSD IOPS (small writes from a database, for instance) into a few dozen HDD IOPS. This allows your storage array to absorb even tens of gigabytes of random writes at hundreds of megabytes per second into the SSDs, which then get reorganized, optimized, and streamed to the HDDs sequentially in a way that optimizes throughput for the HDDs. And program-level calls to sync() or fsync() can legitimately return after completing the writes to the ZIL, even if the writes are still pending to the HDDs, because the data is genuinely on persistent storage that will survive a power outage.

You can also have an L2ARC (Level 2 Adaptive Replacement Cache) with ZFS, which is basically a page cache for *reading* from the HDDs that sits on a partition in the SSDs. For my servers, I set up the L2ARC to consume about 75% of the space of the SSDs because I don't tend to get very large bursty writes on my workload, but for those with a much higher write workload they will want to increase the percentage of ZIL vs. L2ARC.

Once again, like the ZIL, the advantage of L2ARC is to reduce the workload on the HDDs and reduce the IOPS demanded of them. The ARC algorithm has also been mathematically proven to be generally more efficient at common page cache workloads than the page cache algorithm the Linux kernel uses for other filesystems, so there's a "Layer 1 ARC" in RAM, too. And it's adjustable in size so you can tune whether you want to suck up lots of RAM with ARC, or leave more RAM for application data.

For those who would just have HDDs and use RAM buffers to insulate the storage from high IOPS, RAM has three major limitations: one, it's volatile, so it can't safely cache writes for very long; two, using RAM for filesystem caching competes with applications that want to allocate RAM for their own purposes; and three, RAM is very expensive. Also, it's much easier to expand the amount of storage in a server than to expand the amount of RAM: if you have the max. amount of RAM your motherboard supports, you'd have to buy an entire new system to get more. With storage, you can usually just attach another drive, pair of drives, or worst case attach another SATA, SAS, or NVMe card onto the PCIe bus using a spare slot for even more storage. Long story short, you can have a much smaller scale system with terabytes upon terabytes of HDD or even SSD storage, but servers with a couple terabytes of RAM are absolutely enormous, come at a massive cost premium, and require special planning for power, rack space, and system administration, which usually isn't required if you just add a few storage devices.

So, if you want the best of all worlds, being able to use relatively inexpensive commodity hardware (like single-slot Xeon, or even desktop-grade hardware like Threadripper) but have excellent performance for workloads like databases, game servers and so on -- anything that demands a lot of small writes -- your most affordable path is to use Tiered Storage.

You would think that Linux would have a stable, mature, tested, highly optimized filesystem in-house for handling Tiered Storage properly, but it actually doesn't. Not at all. None of the solutions available with Btrfs, XFS, Ext4, LVM2, MD, and family even come close to the performance and feature-set of ZFS with tiered storage. Not to mention that the closest feature competitor, Btrfs, is still such a boondoggle stability-wise that Red Hat is abandoning it as a supported filesystem in RHEL. They also don't have any engineers to work on it, but if it were stable, they wouldn't need to.

I will continue to use ZFS on Linux (at my own peril? Fine.) until Linux offers an in-kernel alternative that matches its performance, featureset and maturity. LLNL has the right idea -- they knew what they were doing when they invested so many dollars into the development of ZoL. They needed a tool that didn't exist, so they built one.

And no, running a Solaris or BSD kernel probably isn't a viable alternative, when most all software is designed and tested for Linux, and the BSD and illumos compatibility layers for Linux are sketchy at best.

For Linux laptops and home gamers, using XFS on a single HDD or SSD is fine, and even if you have a system with both HDDs and SSDs, tiered storage probably isn't of major benefit to you, because you don't have a workload that justifies it. A lot of people do, though, so they're either paying out the nose for way more SSD storage than they need, way more RAM than they need, way bigger servers than they need, ... or they're smart and they use ZoL.

[k1e0x]

Of course when I speak about Users I speak about their own machines / Desktop, I don't go around and say hey "Linux" is on most computers because, it drives the Internet and all 10 quadrilliards of smartphones use it. The desktop is what matters and there Linux has like 2% or something und BSD like 0.1% if not less.

Well.. yeah.. ? FreeBSD is a server OS.. so is Linux really (at least it use to be..) And ZFS is an enterprise filesystem.. It has some cool uses in the home but ZFS runs 100 petabyte arrays, ZFS dosen't compeate with Ext4.. it competes with NetApp's WAFL filesystem. People don't really want it in Linux tree as ryao has said. Most people are pretty happy with where it is.. they just want it to work well with Linux so they can use it to free their enterprise infrastructure from the likes of EMC, NetApp and DDN.

Desktop is.. whatever floats your boat.

[oiaohm]

As far as I know, Oracle was able to settle the lawsuit without giving very much. They did not want to divert resources from the Java lawsuit and Netapp did not want to be in the lawsuit. Netapp did get what they originally wanted though, which was not having to pay royalties over the ZFS patents.

We are talking about legal here. As far as I know is not good enough. We need either Oracle or Netapp to state what was done in that settlement.

There is no court higher than the US Supreme Court in the US according to the US constitution. It’s decisions are final. If you look at “international courts’” decisions, you would find the decisions are non-binding (and are often ignored by the losing parties).

No this is the problem. International court rules on copyright are binding if they happen to be based on a trade agreements as breaching international trade agreements trigger international sanctions. By the way trade agreements signed off by the President of the USA has a higher authority than the US Supreme Court by the constitution. If Oracle if GPLv2 implied patent license gets ruled against in the International court and gets ruled in favour in the US court the result is Oracle can only sell product in the USA until they obey the GPLv2 implied patent license. The section of international law that creates GPLv2 implied patent license in fact automatically triggers International sanctions on the losing party no matter what they judge says.

I will give you that there are many international court decisions that are not binding. Problem the ones that are binding from the international court hit hard. Yes is safer to take on GPLv3 declared patent license than end up in the international court because you are disputing if patents over rule copyright or copyright over rules patents. International court says copyright overrides patents and if you disagree with that point except not being able sell globally.

Please note it would be highly unwise for the US supreme court different to the international court on GPLv2 implied patent license because the worst the international court could do is a full ban/International sanction on exports of a particular class of product from a particular country in the case of GPLv2 this would be all software that you buy with money from the USA. Basically death to the USA software industry.

Basically international court is anywhere between completely toothless to I will kill you and everyone know just because I can. When comes to the implied patent license in copyright is the worst one. Yes it based on something that is a higher authority than the US Supreme court by US Constitution .

https://en.wikipedia.org/wiki/Treaty_Clause

Basically you are forgetting the effects of the Treaty Clause. And part of some treaties is that the case can be heard and rules in the international court with all countries agreeing to enforce the ruling. So technically with treaties over copyright that the USA has agreed to if Oracle was ruled against the USA would have to stop them from exporting and possibly everyone else in the paid for software field

US Supreme court is not always the highest court in the USA. Treaty Clause at times put the international court above the US supreme court. GPLv2 is one of those horrible cases.

We have the SFLC and plenty of companies willing to work together if there is an issue. Why do you think we have OpenZFS?

Problem is you are not really reading what the SFLC people have told you. They gave you a lot of possible untested solutions and also told you without question CDDL and GPLv2 are incompatible.

As for Oracle losing, they would lose if they went after ZFS too. That is what my legal counsel believes anyway.

The reality is you have none of the documents/prior precedents you need to win for sure with CDDL. Has you legal counsel done a proper devil advocate test on your position.

Reality once you do that on CDDL you find yourself in trouble. Devil advocate means you drop all the straw-man arguments that don't have proper backing.

I'm convinced that the reason Oracle hasn't already gone after commercial implementations of OpenZFS already is that they would lose. They pay people 24/7 to look at issues like this.

Really if Oracle currently sued the parties using ZFS and won there would be a good chance Oracle would spend more money fighting the legal case than they would get. So it does not make any sense at this time. This is one thing about Oracle case track record they go after people who they can make a profit from. So it makes perfect sense that Oracle will be sitting and waiting for ZFS.