Approved: Fedora 33 Desktop Variants Defaulting To Btrfs File-System

[kloczek]

because zfs doesn't provide redundancy on single device(btrfs does)
i don't care about license. and i mean it, i'm not freebsd idiot who hates gpl. but i wouldn't use some unsupported out of tree crap and license has indirect relation to that. but even if zfs was relicensed by oracle and merged to linux, i'd have no use for it, because btrfs is clearly superior. just like i have no use for xfs, it's ridiculous to even consider using non-resizeable fs

Of cource it provides "zfs set copies=N <volume>"
https://docs.oracle.com/cd/E19253-01...vpg/index.html
Try to find something like that on any Linux FS

[pal666]

So which one is working at lease the same well as that one which was in ZFS in 2013?

all of them are out-of-band, i.e. you run some program and it does dedup. ones which use btrfs-specific functionality(there are some cross-fs dedupers), are very efficient. btrfs send/receive provides only newly written extents and btrfs keeps all page checksums, so there's no need to do rescans or checksum calculations etc

[pal666]

I don't understand.. why can't you have ZFS on a single device?

because zfs doesn't provide redundancy on single device(btrfs does)

Pal666... Lets just be honest here.. you just don't like the licence right?

i don't care about license. and i mean it, i'm not freebsd idiot who hates gpl. but i wouldn't use some unsupported out of tree crap and license has indirect relation to that. but even if zfs was relicensed by oracle and merged to linux, i'd have no use for it, because btrfs is clearly superior. just like i have no use for xfs, it's ridiculous to even consider using non-resizeable fs

[kloczek]

there are many different implementations of dedup for btrfs with different properties and overheads. you would know that if you weren't mindlessly reposting zfs propaganda

So which one is working at lease the same well as that one which was in ZFS in 2013?
PS. I'm not asking about the same quality/performance/resources consumption as it is now in Solaris 11.4 last ASRU.

[k1e0x]

no
everyone needs fsck
btrfs also has checksum. and btrfs has another sources even on single drive.

you fail to read. btrfs clones are first class citizens, you can clone from clone. on zfs you can't
btrfs dedup is not set in stone either
smartphones still have only one drive

I don't understand.. why can't you have ZFS on a single device?

pool: zpool
state: ONLINE
scan: none requested
config:

NAME STATE READ WRITE CKSUM
zpool ONLINE 0 0 0
ada0p2 ONLINE 0 0 0

errors: No known data errors

Looks fine to me. Realistically you want something that is as paranoid about data integrity as ZFS on the most flaky of drives.. such as usb sticks. it's good there, it's also cross platform. (linux, bsd, macos, solaris, windows etc)

And you can clone a clone, a clone is just a writable snapshot. btw, ZFS also has another level of "snapshot" called a bookmark where it can undo everything, even deleted volumes (datasets)

Pal666... Lets just be honest here.. you just don't like the licence right? Because you head it was "anti-linux" or something? It's not, the FSF was wrong about it. ZFS is a benefit to Linux because it allows it to do a lot of great things. It allows Linux to compete with NetApp and EMC... I'll bet there are hundreds of people reading this who run Linux VM's backed by NetApp.. wouldn't it be cool if they ran Linux VM's on Linux? ZFS allows us to do this today, the capabilities for enterprise storage are already there.

Let the past die, it was a lie.

[pal666]

zfs works on a single drive

"requires reinstall from backups on every bad block" rather than "works"

, zfs is quite space efficient as well

it is explicitly less efficient than btrfs

, zfs has checksum on every single block in the tree,

that's one checksum per 128kb block instead of 32 checksums of btrfs

zfs can clone from a clone

i trust more that paper than you

, zfs has raid levels and combinations of nesting that we don't even have RAID-X numbers for

btrfs raid 0 and 1 also don't have numbers, i use such numbers only because they are similar, but they are superior

, and zfs can be resized as well (grown).

and people can fly(down)

: shrug : maybe you need to learn a bit more before you attack people?

i'll wait for you

[pal666]

truth is hard. Do the hard work Linux! Engineer, be exceptional. I want that, I would love to see linux change it's philosophy and shine.

just look around. linux's shine dominates. you have to be blind to not see it

[pal666]

That's biased.

no

ZFS does not need fsck and having it would not help it do anything better.

everyone needs fsck

It has no journal to replay but does have the checksum it validates on each read. if it's wrong it will automatically repair (if another source exists)

btrfs also has checksum. and btrfs has another sources even on single drive.

ZFS also supports clones and writable snapshots. BTRFS isn't unique here.

you fail to read. btrfs clones are first class citizens, you can clone from clone. on zfs you can't

ZFS's deduplication is being rewritten to be ram friendly.

btrfs dedup is not set in stone either

Where it is true ZFS was designed for the enterprise and large storage. As we often find in computing what was once only available to
supercomputers is, given time, in every smart phone. Storage data management of our personal systems is looking closer to what enterprise was using a few decades ago.

smartphones still have only one drive

[NotMine999]

buy larger drives, or wipe and reinstall
you can't come up with smarter questions?

After reading the bug report, that I doubt you read, that Fedora referenced I came to the conclusion that they are trying find simple & easy ways for users to manage their systems without forcing the user through complicated steps. To me that sounds a lot like WINDOWS, but I doubt you have ever used WINDOWS or it's SERVER flavors. That is why I made references to "What do Windows users do".

Your comment suggests that Fedora WANTS users to work through complicated steps to manage storage space within the system, but the link in the article and the Fedora bug report referenced by the page from that link don't support your view.

From what I can tell:
Fedora wants their system administration to be easier, so users don't have to follow complicated steps to use all available storage space currently in the system; sounds like a noble goal to me. The linked Fedora page, that I doubt you read, focused on a discussion of methods to have 2 different mount points built on unique partitions and thus having a fixed size (a byproduct of their installer, IIRC) both access the same remaining "free space" in the system, and it looks like they figured out a way to do it. For some users that will solve some problems. It does not yet appear to solve the inevitable problem of actually running out of storage space on existing media where the only solution left is adding another drive and possibly moving 1 or more mount points and their associated files to the new larger storage media; a complex and complicated task that newbies would certainly mess up (a goal Fedora is trying to avoid) and experienced users generally groan about while accomplishing the task.

[k1e0x]

btrfs works on single drive, btrfs uses btree of extends which is more space and time efficient, btrfs checksums individual pages, btrfs can clone from clone, btrfs has raid 1 and 0. and not on that list, but still important: btrfs can be resized

zfs works on a single drive, zfs is quite space efficient as well, zfs has checksum on every single block in the tree, zfs can clone from a clone, zfs has raid levels and combinations of nesting that we don't even have RAID-X numbers for, and zfs can be resized as well (grown).

: shrug : maybe you need to learn a bit more before you attack people?