Red Hat's Stratis Storage 3.0 Released With Many Improvements

[cynic]

You certainly have cogent reasons to think so. Care to share them?

nope!
when someone starts a post with "I believe" it is not meant to be taken as if it were an undisputable universal truth that requires incontrovertible proofs.

[Old Grouch]

You know, with a COW filesystem, you should be able to take any file and scroll backwards and forwards write by write giving you 'infinite' file versioning. Nothing that I know of gives that, not even NILFS2 (although the underlying model allows it). Combined with checksumming and/or erasure codes, you'd have a pretty good data store. As far as I know, ZFS, btrfs, bcachefs, NILFS2, F2FS+LVM or anything else doesn't make this possible. Which is a shame.

[vb_linux]

https://pkgs.org/download/stratisd

Not 100 percent right. arch has picked up stratis. Big barrier is debian has not picked stratis up. Debian does have quite a large market share effect.

Good to see all the Ubuntu haters coming out and attacking Red Hat for "Not Invented Here" syndrome. If it was Ubuntu, everyone would have been shouting that why can't they invest and work with btrfs guys.

[oiaohm]

Good to see all the Ubuntu haters coming out and attacking Red Hat for "Not Invented Here" syndrome. If it was Ubuntu, everyone would have been shouting that why can't they invest and work with btrfs guys.

https://stratis-storage.github.io/

There is a catch stratis storage is unified management tool. Stratis takes the existing XFS file system and the Linux kernel device mapper system and takes this as far as it can go.

Nothing stratis does you could not do on a Ubuntu system with other management software.

So this is a double "not invented here" syndrome question. btrfs and zfs both remake large sections device mapper layer in their file system drivers so both are gulity of doing a reinvent of well. stratis is example that that question if this revent by btrfs and zfs is wrong but there is another bcachefs as well. Yes bcachefs does use the device mapper layer on Linux.

Stratis might be the wrong answer long term but btrfs and zfs could be the wrong long term answer as well.

Yes I know it appears to be double standard but the hard reality here is when people have yelled at ubuntu for reinventing wheel the developers normally goes around claiming a stack of fake limitations why their solution should be pushed. The old saying if you are standing in a glass house don't throw stones. Yes as horrible as it sounds developers around ubuntu getting yelled at for NIH is normally because they have stupidly thrown the first stone by claiming their solution is superior with some false fact.

Redhat has not gone around claiming any false facts about stratis attempting to drum up market share. Yes you can have NIH syndrome in the open source world and get away with it. The critical requirement to get away with NIH syndrome without being yelled at is don't claim anything false as a reason why someone should use what you are doing.

[vb_linux]

I always assume a lot of people get paid by Redhat/IBM directly or indirectly here. So, even when Ubuntu was the only one trying to push Desktop Ubuntu hatred was spread against them like why a small company was not contributing the same amount as billion-dollar contribution. Lol! really, while secretly afraid of the popularity of Ubuntu amongst masses. Now that the Ubuntu desktop dream is dead. Enjoy.

Yes I know it appears to be double standard but the hard reality here is when people have yelled at ubuntu for reinventing wheel the developers normally goes around claiming a stack of fake limitations why their solution should be pushed. The old saying if you are standing in a glass house don't throw stones. Yes as horrible as it sounds developers around ubuntu getting yelled at for NIH is normally because they have stupidly thrown the first stone by claiming their solution is superior with some false fact.

And the biases in your post are so obvious.

Which developer would start a project with the aim of it being worse than the current standard?

Thing People here think the source is open so anyone can use it. Not so, in a big project. The support money and contracts go to companies who own the major projects. Company which has the coders working on the big project and understand that project.

That is the sole and only reason Redhat does not want to push btrfs or try to improve it to a point where it is the excellent and best solution. Redhat owns gnome, and all the bile against unity was not because of false facts but because of money and controlling the major UI platform on linux.

Redhat has not gone around claiming any false facts about stratis attempting to drum up market share. Yes you can have NIH syndrome in the open-source world and get away with it. The critical requirement to get away with NIH syndrome without being yelled at is don't claim anything false as a reason why someone should use what you are doing.

NIH syndromes are good up to 2-3 competitors. Beyond that is confusion. Projects become stale, too cocky, developers develop egos and in open source you can't fire them. But you can fork, and one should.

[oiaohm]

I always assume a lot of people get paid by Redhat/IBM directly or indirectly here. So, even when Ubuntu was the only one trying to push Desktop Ubuntu hatred was spread against them like why a small company was not contributing the same amount as billion-dollar contribution. Lol! really, while secretly afraid of the popularity of Ubuntu amongst masses. Now that the Ubuntu desktop dream is dead. Enjoy.

No this is not the problem.

And the biases in your post are so obvious.

Which developer would start a project with the aim of it being worse than the current standard?

There is a different between aiming to be better than the current standard. And what ubuntu developers have done many times.

NIH syndromes are good up to 2-3 competitors. Beyond that is confusion. Projects become stale, too cocky, developers develop egos and in open source you can't fire them. But you can fork, and one should.

https://www.linuxvoice.com/interview-thomas-vos-of-mir/

Thomas: Right. I find Wayland impressive, obviously, but I think Mir will be significantly more relevant than Wayland in two years time.

Remember this is the lead developer of Mir.

No one was buying into that; no one was saying, “Look, we’re moving this to production-level quality with a bona fide protocol layer that is frozen and stable for a specific version that caters to application authors”.

Then this line. Guess what there were people in 2014 buying into the wayland idea with that idea. So Thomas here has basically slapped them straight across the face. Yes its a false claim. At this point you have a problem you have many upset people.

I could pull the same thing out with Unity and other Ubuntu started projects that died under Ubuntu management.

vb_linux like it or not sometimes is not best move to let the lead developer have a interview without supervision. Ego and cocky in actions in media are very quick ways to get other peoples back up and end up having project tared with this project is just someone suffering from NIH.

Thing People here think the source is open so anyone can use it. Not so, in a big project. The support money and contracts go to companies who own the major projects. Company which has the coders working on the big project and understand that project.

That is the sole and only reason Redhat does not want to push btrfs or try to improve it to a point where it is the excellent and best solution.

https://en.wikipedia.org/wiki/XFS No it is not the sole and only reason. It really simple to forget Redhat has developed XFS for decades.
Why should not lead developer of XFS not improve XFS to the point where it is the excellent and best solution. This is what stratis-storage is to take XFS as far it can go.

Of course BTRFS and ZFS developers should have been planning for the day they would have to take on the best of the existing solutions. Redhat with straits storage is really just setting a min bar.

The reality 98% of what straits storage is happens to be automation of what was in the Red-hat administrator guides of the 2000s. Staitis storage is basically the existing solution streamlined. The pre-existing solution is what btrfs, zfs and bcachefs should be able to beat by a large margin. Problem btrfs and zfs don't beat stratis by a large margin in fact in many cases loses to stratis.

https://lwn.net/Articles/867427/

This also explains why redhat has kind of run away from btrfs. Do notice everything else is under 3G of ram and btrfs magically blows out to 16G of ram. Yes ZFS also surfers from this problem.

Staitis storage does not eat you out of memory and have degrading performance like ZFS and BTRFS both can suffer from.

There is a possibility that BTRFS and ZFS are both not fixable. Yes this could be why bcachefs in time beats them.

Which developer would start a project with the aim of it being worse than the current standard?

There is a side to this. A developer may not intentionally start of with the aim of being worse than the current standard. BTRFS and ZFS seam to show that is possible totally by accident to aim to that. Yes without Stratis Storage making the existing easier to setup you don't have something quickly bench BTRFS and ZFS and bcachefs against to really see if they are improvement.

Lot of ways redhat working on .Stratis Storage make the current standard before BTRFS/ZFS/bcachefs as good as it can be that people developing BTRFS/ZFS/bcachefs if they are in fact aiming to make something better than the current standard. Or if they are in fact aiming to make something worse than the current standard.

Stratis Storage is not really NIH. Stratis Storage more something that should have been done before we started BTRFS, ZFS and bcachefs so we had a solid base line to work from so developers knew if the solutions they were making were better than existing or worse. Better late than never.