Red Hat Appears To Be Abandoning Their Btrfs Hopes

[Steffo]

I'm pretty sure I sensed innuendo in there that Redhat was expressing interest in a new filesystem with the same intentions btrfs had. They more or less asked their customers to give them feedback on desires for features. That's where I think the innuendo is.

Now it makes sense! Thx!

[starshipeleven]

Oracle is a purely self-serving, money-grubbing company. But they gain no financial, legal, public relations, or other benefit by making btrfs awful.

Oracle has some obvious interest in btrfs not sucking. They are offering products where it is a main selling point, namely Oracle Linux. http://www.oracle.com/technetwork/se...ies/index.html
And the fact that btrfs isn't Oracle's own stuff prevents them from running into the ground, which is a plus.

[Steffo]

Maybe the problem is, that till today, Btrfs is still not completely stable (RAID 5/6) and it is not a general file system! It is not suited for databases or VMs, because the performance sucks on this scenario, because of COW. But if Red Hat really develops an own next generation file system, wouldn't it have similar problems?

[starshipeleven]

Nonsense, when you have a paid support contract from a vendor, Red Hat for example, they are on the hook for the quality of the product - that's what you are paying them for. That's what "support" means. It means that when it breaks, they are obligated to help you fix it.

I'm actually saying the same thing. Drivers per-se don't give any warranties be they open or closed. It's companies that decide if it is worth to take their odds at sellign you a support contract over some technology or not.
It seems RedHat does not see profit in offering support for Btrfs, but that guy's claims that the drivers could offer some warranties by themselves is nonsense.

I mean, SUSE is supporting Btrfs fine, the world is not ended yet.

Why are you trying to use this worldwide de facto standard of limited liability, and applying it to open source software as if that's a unique use case?

That was not my intention, I admit I could have been more clear that limited liability wasn't limited to Linux.

[schmidtbag]

Here, read NVIDIA's blob license:

That doesn't explicitly say all their drivers are betas or to use them at your own risk. What they're saying is to use the software they want you to use, and that you can't sue them in the event something breaks. That doesn't mean they're not stable or reliable, it's just a way for them to not get sued when you configure something wrong.

And here a HP driver license (printer drivers, I don't know if they are for Linux or not) https://support.hp.com/us-en/document/c00581401

Disclaimer 9 supports your point (seriously, the viruses part is a little scary), except the last sentence which basically forces HP to take responsibility for their faults, depending on where you're from. Disclaimer 10 is the same thing as 6.2 of Nvidia's - it's to protect them from getting sued, but doesn't say anything about the stability or reliability of the product.

Huh? It's not like they don't have full control of what goes into their kernels didn't they?

True, but I don't blame them for not wanting to deal with it. The average alligator can run at 20MPH (32KMH), but chasing after a gopher in the distance takes too much energy for such a small yield; it'd rather just sit there and wait to be approached. Btrfs is no different - it "moves" too much for how little substance RH gets out of it. I figure they'll reconsider adding it again when performance stabilizes.

Yeah, also Facebook, and that evil monster Fujitsu. Man I'm sure I've also seen Chtulhu posting patches in the mailing list.

Your attempt to undermine the significant of my comment accomplished nothing. The fact of the matter is, there's a reason why forks like LibreOffice, OpenIndiana, and MariaDB have come about - Oracle does not make for a good parent. I'm not calling them evil, but it is well-known they can make things difficult for many people who use their software.

Knowing RedHat, it's probably because their average customer uses databases that have checksums already, or expendable webservers where it's irrelevant anyway.

That only addresses one of many beneficial features Btrfs has. As far as I'm aware, there is no complete Linux-native replacement to it.

[carewolf]

I've been using BTRFS for going on 6 years now, as single devices (for corruption detection) and as RAID1 for availability on multiple systems and drives. It has saved me from a failed drive controller that was corrupting data it wrote to the platters, and it has never failed me. Given the option, I wouldn't go back to EXT4/LVM2/mdraid. If it gets pulled from Fedora, I'll switch to another distribution before switching to another filesystem.

I have used it twice, and both times it irrovokably bitrotted the disk, and with no proper fsck, and a filesystem that refuses to read faulty metadata it created itself, it just eats and destroys data. Really tragic they still havent made a proper fsck.

[carewolf]

FYI: all Linux drivers are "use at your own risk", so is most opensource stuff.
It's not like you can go and sue ext4 devs if your filesystem fucks up and eats your data, or AMD/NVIDIA if your GPU drivers crash and you lose a competitive Overwatch match for world championships or something.

The difference is that when data goes bad on ext4 filesystems, you lose a file or a bit of a file, when something goes wrong with btrfs you lose the entire partition.

[starshipeleven]

That doesn't explicitly say all their drivers are betas or to use them at your own risk.

Dunno, I thought that an explicit "provided AS IS", giving "no warranties" and "decline any liability" meant that you use them at your own risk in English.

I'm not discussing the fact that their drivers may be good or not. Just that all drivers usually are provided "at your own risk", and no driver ever exits from that stage. You can have drivers or features that are assumed stable and reliable, but not official legal binding statements that they are.

doesn't say anything about the stability or reliability of the product.

Dunno, I can only say I'm not very swayed to believe in their stuff after they washed their hands by providing it as-is and giving no warranties and declining any liability.

Your attempt to undermine the significant of my comment accomplished nothing.

My point was that btrfs isn't "their software" any more than it is fujitsu's or SUSE's or Facebook's (or Chtulhu's). That's my point. Oracle is just a major contributor, not an owner. They can't decide btrfs fate on their own.

That only addresses one of many beneficial features Btrfs has. As far as I'm aware, there is no complete Linux-native replacement to it.

Yeah, but RedHat isn't a charity, if their customers don't seem to need it, then why should they invest in it. I assume that they know how to businness like this

[starshipeleven]

The difference is that when data goes bad on ext4 filesystems, you lose a file or a bit of a file, when something goes wrong with btrfs you lose the entire partition.

ext4 can refuse mounting too

https://bbs.archlinux.org/viewtopic.php?id=143266

https://superuser.com/questions/5757...perblock-found

EDIT: for the sake of fairness, also XFS can refuse mounting

https://superuser.com/questions/575783/cant-mount-ext4-ssd-jbd-no-valid-journal-superblock-found

https://askubuntu.com/questions/486093/cant-mount-xfs-filesystem-after-reboot

https://superuser.com/questions/384753/xfs-cant-read-the-superblock

[starshipeleven]

I have used it twice, and both times it irrovokably bitrotted the disk, and with no proper fsck, and a filesystem that refuses to read faulty metadata it created itself, it just eats and destroys data. Really tragic they still havent made a proper fsck.

Please state dates or versions, and note that in the last years the fsck and scrub improved significantly.