Tuxera Claims NTFS Is The Fastest File-System For Linux

[Apopas]

current linux is far from perfect .

So is windows and any other human creation...

[deanjo]

So is windows and any other human creation...

Lies! You have not tried my chicken noodle soup. NO SOUP FOR YOU!

[locovaca]

ECC memory is supported on all AMD CPUs since socket 754 days, and only recently AMD started to screw consumers by dropping it from their Fusion parts. I think the majority of AM2/AM3/+ mobos support it too.

Small correction, Socket A Athlon MPs required ECC in some configurations.

[kebabbert]

If the problem is resolved in the block layer it should probably make every native Linux file system safe from the silent data corruption.

There is a big difference between theory and practice. Even if what you claim is true in theory, there might be problems in Oracle's implementation. Thus, we need someone else, preferably researchers, that examines the Oracle solution.

You know, in theory, RSA cipher is considered as quite safe, but in practice there might be problems in the implementation of RSA that an attacker can use.

Like before: there were some changes aimed at the issue. Check this:

http://blogs.oracle.com/linux/entry/...ata_corruption

There's a white paper about sdc and Linux.

I like this part.

Yes, but white papers are not to be trusted, you have said to me many times. I showed you white papers earlier, but you rejected them all. You said that the white papers that SAP created at www.sap.com, is Sun FUD and Sun propaganda. You said that SAP is business partners with Sun and therefore you can not trust on the official SAP white papers. The odd thing is that SAP is also business partners with several Linux companies: RedHat, SuSE, etc. So SAP should favour Linux, not Solaris.

When I show white papers to Kraftman, they are FUD and lies. When Kraftman shows me white papers, they are good and to be trusted.

Anyway, I prefer research instead of white papers. The best would be if different research groups established the same conclusion - but you can not get everything.

Until there is research on Oracle Unbreakable Linux Kernel and Silent Data Corruption, I would not by sure that the Oracle solution is safe. Maybe research will show Oracle's solution is unsafe, just like everything else: NTFS, ext3, ReiserFS, XFS, JFS, Hardware Raid, ...

[kraftman]

Yes, but white papers are not to be trusted, you have said to me many times. I showed you white papers earlier, but you rejected them all. You said that the white papers that SAP created at www.sap.com, is Sun FUD and Sun propaganda. You said that SAP is business partners with Sun and therefore you can not trust on the official SAP white papers. The odd thing is that SAP is also business partners with several Linux companies: RedHat, SuSE, etc. So SAP should favour Linux, not Solaris.

Afaik I said sap sucks. SAP was open source hostile those times, but times changes. However, maybe they're still hostile? And yes, SUN spreaded FUD all the time, but thankfully SUN's dead.

When I show white papers to Kraftman, they are FUD and lies. When Kraftman shows me white papers, they are good and to be trusted.

They're from your favourite source, so you don't trust your favourite source? No, the lies and FUD were yours and SUN. You were giving outdated papers usually or not related to disscusion.

Until there is research on Oracle Unbreakable Linux Kernel and Silent Data Corruption, I would not by sure that the Oracle solution is safe. Maybe research will show Oracle's solution is unsafe, just like everything else: NTFS, ext3, ReiserFS, XFS, JFS, Hardware Raid, ...

Don't care. Until there there's no present research on ZFS I'm not sure it's still SDC safe. You know, bugs happen and while Oracle leads the project now, they could mess it up or something. Do you have some current research on Oracle's ZFS?

[kraftman]

I wouldn't trust ZFS, because of such bugs:

http://blog.lastinfirstout.net/2010/...oracle-on.html

http://blog.lastinfirstout.net/2010/...oss-still.html

There are many problems with ZFS, so I prefer to use Ext4.

[Apopas]

Lies! You have not tried my chicken noodle soup. NO SOUP FOR YOU!

Heh I'm lucky I don't like noodles then

[kebabbert]

I wouldn't trust ZFS, because of such bugs:

http://blog.lastinfirstout.net/2010/...oracle-on.html

http://blog.lastinfirstout.net/2010/...oss-still.html

There are many problems with ZFS, so I prefer to use Ext4.

Yes, there are some problems with ZFS, as it is a young filesystem. It takes decades to iron out all bugs. New code always has bugs. ZFS code is from 2004, so it is only 7 years old. Many Enterprise sysadmins refuse to use filesystems that are young (less than a decade). So... how long do you think it will take before sysadmins will use BTRFS v1.0? BTRFS v1.0 is many years away. After v1.0, serious sysadmins will wait another 5-10 years before they deploy BTRFS. BTRFS is not production ready and will not be in 10 years.

Regarding ZFS bugs, yes there are ZFS bugs. There are bugs in every piece of software. If you believe ext4 has no bugs, I suggest you look at this site:
https://bugzilla.kernel.org/buglist....component=ext4
Some ext4 bugs is about data loss, too.

The difference between ZFS and ext4, is that ZFS is built up from ground with a focus on protecting your data against Silent Corruption. No other filesystems are targeting Silent Corruption protection. There might be bugs in ZFS, yes. But when ZFS does not run into bugs, it protects your data according to research.

ext4 might also run into bugs so it will corrupt your data (see the list above). But even when ext4 does not run into bugs and ext4 functions exactly as intended, ext4 will not protect your data. Because ext4 is not designed to protect your data as extensively as ZFS. All filesystems tries to protect your data. Research shows that only ZFS gives good protection. Research shows that all other filesystems fail.

I suggest you continue to trust ext4, and I will continue to trust on ZFS. Dont hold your breath for BTRFS though.

[kraftman]

If you believe ext4 has no bugs, I suggest you look at this site:
https://bugzilla.kernel.org/buglist....component=ext4
Some ext4 bugs is about data loss, too.

No, I agree there are bugs in every piece of software. I consider Ext4 to be enough for desktops and it's probably enough for many other areas.

The difference between ZFS and ext4, is that ZFS is built up from ground with a focus on protecting your data against Silent Corruption. No other filesystems are targeting Silent Corruption protection. There might be bugs in ZFS, yes. But when ZFS does not run into bugs, it protects your data according to research.

btrfs is also targeting at this.

ext4 might also run into bugs so it will corrupt your data (see the list above). But even when ext4 does not run into bugs and ext4 functions exactly as intended, ext4 will not protect your data. Because ext4 is not designed to protect your data as extensively as ZFS. All filesystems tries to protect your data. Research shows that only ZFS gives good protection. Research shows that all other filesystems fail.

Research is old. Linux + Oracle's and their friends' infrastructure should give you even better protection than ZFS.

[kebabbert]

Afaik I said sap sucks. SAP was open source hostile those times, but times changes. However, maybe they're still hostile? And yes, SUN spreaded FUD all the time, but thankfully SUN's dead.

We had this discussion many times before. You still dont get it, do you?

So what is FUD? It is basically, a negative lie. The official SAP benchmarks with Linux and Solaris, are no lies. Anyone can replicate the numbers. SUN did not say "we got 100.000.000 in SAP benchmarks" - Sun did not lie. SUN claimed they got a certain benchmark score, and SUN did get it. SUN did not lie. But you lied. You quoted wikipedia on "FUD" and you confessed that you FUD sometimes. Yes, you wrote that. Kraftman is a FUDer. I can link to your post where you confess that you FUD.

So, again: SUN did not lie. It was a controlled benchmark from SAP. And SAP controlled the Linux benchmarks too. Yes, I know you dont like that Linux lost the benchmark, but that is the way it is. Sorry. SAP did not FUD. It is you who dont understand what FUD is, even though you read and quoted wikipedia on FUD, you dont understand what you read.

They're from your favourite source, so you don't trust your favourite source? No, the lies and FUD were yours and SUN. You were giving outdated papers usually or not related to disscusion.

I gave papers that was a few years old, yes. But you have never showed any papers. You have claimed lots of things, most of it were lies. For instance, you claimed that Solaris scales to 64 cpus. People posted links to Solaris servers with more cpus, I posted a link to a Solaris server with 144 cpus. But still you lied about Solaris scales only 64 cpus, even though you saw evidence of Solaris servers with more cpus. That makes you a liar and FUDer. And you never gave any papers showing that Solaris only scales to 64 cpus. The reason you did not give such papers: they dont exist. Solaris scales much higher.

I have many times asked you to post any papers that support your claims, but you never did. Never ever. Not a single paper. I have posted several papers, that was 2-3 years old, yes. You never posted any papers.

Don't care. Until there there's no present research on ZFS I'm not sure it's still SDC safe. You know, bugs happen and while Oracle leads the project now, they could mess it up or something. Do you have some current research on Oracle's ZFS?

I have posted research paper on ZFS and Silent Data Corruption above here, which you know I have. I have showed you that paper several times before, but you denied it to exist. You did not reject the research, but you denied the paper existed. I posted the paper, and you said something like "Liar, show me the paper! Where is the paper??" And I posted the paper again, and you kept repeating like a robot "where is the paper?? Show it to me!!!"

So, I dont think there is any use if I post the same paper again here. You will probably deny it's existence. You will pretend to not see the paper, no matter how many times I show it to you. And at the end you will probably say something like: "haha, you did not show any research paper on ZFS and SDC, which means you lie! FUDer! Moahahahaha!!! Where is the paper??? I can not see any paper, you are a FUDer!!!". Just like you did the previous time we had this same discussion.