Hello, the btrfs is still experimental and under heavy development, until it becomes stable, we should only provide it as an expert option.

Status: NEW => RESOLVED
CC: (none) => shikamaru
Resolution: (none) => WONTFIX

dude, that's totally not true, btrfs has been disk-stable since iirc 2.6.27 and included in mainline kernel...

there's even distro's rolling it out as default one...

imho don't go off deciding this on your own, let's ask someone with some good kernel knowledge like pterjan or tmb...

Status: RESOLVED => REOPENED
Resolution: WONTFIX => (none)

(In reply to comment #2)
> dude, that's totally not true, btrfs has been disk-stable since iirc 2.6.27 and
> included in mainline kernel...
> 
> there's even distro's rolling it out as default one...
> 
> imho don't go off deciding this on your own, let's ask someone with some good
> kernel knowledge like pterjan or tmb...

Good point; I assume you know how to use the CC function in bugzilla? no? OK, I'll CC tmb myself... (I don't know pterjan's email in LDAP, so can't CC him).

(FWIW, there's always an implicit AFAIK, just sometimes people seem to miss it).

CC: (none) => tmb

No, thatâs not right, from btrfs wiki (https://btrfs.wiki.kernel.org/index.php/Main_Page):

âBtrfs is under heavy development, but every effort is being made to keep the filesystem stable and fast. As of 2.6.31, we only plan to make forward compatible disk format changes, and many users have been experimenting with Btrfs on their systems with good results. Please email the Btrfs mailing list if you have any problems or questions while using Btrfs.

Note that Btrfs does not yet have a fsck tool that can fix errors. While Btrfs is stable on a stable machine, it is currently possible to corrupt a filesystem irrecoverably if your machine crashes or loses power on disks that don't handle flush requests correctly. This will be fixed when the fsck tool is ready. â

Many unstable things get into mainline kernel, that does not mean they are ready for production, and I bet btrfs driver is marked as experimental. Could you point me to a distribution that goes with btrfs by default ? AFAIK meego go with that, but their stable release isnât expected too come out that soon and is still under heavy development, links to prove what you say would be welcome.

As far as Iâm concerned, I prefer to go better safe than sorry, remember that the goal for first release is not to bring new features.

btrfs got included upstream in 2.6.29-rc1

As of kernel 2.6.37 I do consider btrfs stable for usage, 
(it would have been 2.6.36, but due to a late git pull request to Linus, it got postponed)

but the issues are:
- missing error fixing tool
  (btrfsck finds the errors, but can not repair them yet)
- no grub/lilo support
  (can be worked around with separate /boot)


So for now I think it should stay in "expert" mode, because of the missing error fixing.... 

This can be re-evaluted if the btrfs-progs gets their needed fixes soon
(Chris Mason (btrfs maintainer) stated in Q3/2010 that he probably would work on in in the coming months, but so far nothing have happend)

@ahmad: i'm sorry about the implicit AFAIK, but tbf, it did look like a bit like an uninformed drop.

@shikamaru: that wiki page is mostly unchanged since around 2.6.27 , there is a fsck tool as tmb pointed out. (and i was talking about meego, and tbf i wouldn't place it as default at all)
Also, if jfs, xfs and reiserfs is in the same list, ... even though some of those are already present for a long time, it doesn't mean they have no issues (i still hear of failures in reiserfs and xfs, and i wouldn't trust jfs with my /tmp . don't forget that last year ext4 "failed" as well.

But i understand your point

@tmb: thanks for your insight and extra info, imo, we should follow your judgement on this. perhaps we should wait until both of the issues are resolved.

(In reply to comment #6)
> @ahmad: i'm sorry about the implicit AFAIK, but tbf, it did look like a bit
> like an uninformed drop.
> 

You got me wrong, I meant there's an implicit AFAIK in shikamaru's post, that's what the btrfs wiki page on kernel.org said, and he's not a kernel hacker.

[....]

> @tmb: thanks for your insight and extra info, imo, we should follow your
> judgement on this. perhaps we should wait until both of the issues are
> resolved.

There's no "perhaps", we must follow his lead since he's the kernel maintainer/hacker :)

(In reply to comment #7)
> You got me wrong, I meant there's an implicit AFAIK in shikamaru's post, that's
> what the btrfs wiki page on kernel.org said, and he's not a kernel hacker.

ah, good point

> There's no "perhaps", we must follow his lead since he's the kernel
> maintainer/hacker :)

it isn't because we have no choice, that we shouldn't try to phrase this as if we did have a choice :-)

and otoh, the part i was talking about, is about the list of available options in non-expert mode, which is not in the kernel package, but is in /soft somewhere, iirc. :-P

but ok, you can close it as appropriate.

Some updates...

~2 weeks ago Chris Mason stated the new btrfsck is 90% ready and he's working fulltime on it now, so if we are lucky, it will be ready at the same time we switch to 2.6.38 kernel wich has seen some nice improvements for btrfs :)

awesome news!

it appears grub 1.99 now supports btrfs ; is this something that can be patched into 0.9x version?

Did the problem still exist in Mageia 1 rc? (And in Mageia 1?)

CC: (none) => m.van.waes
Source RPM: (none) => drakx-installer-stage2
Whiteboard: (none) => NEEDINFO

Yes I think.

CC: (none) => thierry.vignaud
Whiteboard: NEEDINFO => (none)

See comment 5, 9 and 11. There could be a way to do this for cauldron (maybe). I guess it would be up to the btrfs maintainer to decide...

(In reply to comment #14)
> See comment 5, 9 and 11. There could be a way to do this for cauldron (maybe).
> I guess it would be up to the btrfs maintainer to decide...

There isn't a btrfs maintainer yet :(

Well,

the btrfs-progs that could fix any problems on filesystem level are still not available.


Last stats update by upstream on August 3, 2011:

"This part is definitely my fault.  I've gone through a bunch of
variations on bigger and smaller tools, and had to juggle the kernel
maintenance as well.

Aside from making sure the kernel code is stable, btrfsck is all I'm
working on right now.  I do expect a release in the next two weeks that
can recover your data (and many others)."


But still nothing...

that is too bad... :-( in the main time perhaps you could be a btrfs maintainer?

Added UPSTREAM keyword

Last message (oct. 10th) of Chris Mason I found in http://www.mail-archive.com/linux-btrfs@vger.kernel.org/msg12717.html:

I wish there were bigger forces at play here than me just not having it
finished yet, but that's all there is to it.

Keywords: (none) => UPSTREAM
Source RPM: drakx-installer-stage2 => drakx-installer-stage2, btrfs-progs

Yeah,
I'm monitoring the btrfs ml pretty much on a daily basis in hope of working tools ...

But at the rate it's going, I'm not even sure we will have in Mageia 2

that's too bad, i hope chris has some time to finish this, so we can get started with it for real.

Well, we finally have some movement..

Chris have now merged Hugo's integration tree (wich is mostly bugfixes / enhancements to current code)

Next up for merge is Josef's recovery tool, and then Chris metadata scanning code will be merged and a new beta will be released to testing.

I just pushed an updated snapshot with the integration tree fixes to Cauldron

awesome!

@ tmb

any more news?

Please assign to yourself if upstream has come with a patch. (And then please set to the assigned status too, or, if for workflow reasons you can't do that, put OK on the whiteboard instead).

last snapshot updated to was of 2011-12-01, on december 24th and tmb did another update shortly after that.

@ tmb

assigning to you now, because you clearly showed your intention to import the fixed package when it is available, and that'll be soon, if it didn't already happen :)

Please set status to assigned when you are working on this bug

Assignee: bugsquad => tmb

Yep.

We are working having btrfs available as an option for Mageia 2, but it wont be used by default.

Upstream says final btrfs-progs rewrite should be available around Feb 15, 2012 (but I'll belive that when I see it)

Status: REOPENED => ASSIGNED
Depends on: (none) => 4238

Time to package a newer snapshot?
(See http://www.phoronix.com/scan.php?page=news_item&px=MTA2MDI)

hmm, still, perhaps wait a bit longer... allthough obviously this could be oracle delaying it so it can have it first on their own linux...

Or more likeley they rushed in order to include it but eventually it's not that good near/past the deadline and they're painfully working on it...

afaik only cmason is working on the fsck tool. and he was "nearly" finish already some time ago, which means oracle is giving him other stuff to do...

Yeah, it was supposed to be released in final version around February 15th, but as usual nothing :/

I've been monitoring that "dangerdontuse" branch, and will package a newer snapshot soon, but probably not before beta2 is out...

btrfs is available

Resolution: (none) => FIXED
Status: ASSIGNED => RESOLVED

yay :-)

is it now show on the standard stuff? or do i still need to select advanced?

i'm guessing having btrfs in the standard stuff is still offlimits?, perhaps we can put this on mga4 milestone if necessary

It should already be available providing you switch to expert mode.
Can you check that?
I did quite a lot of fixes for btrfs support but I don't remember all the details.
We could enable it in std mode if fsck works fine (Thomas?).

Keywords: (none) => NEEDINFO

yep it's already in expert mode

for mga2 i already used it with expert mode

In advanced partition mode you can add btrfs partitions, however the module is not loaded so mounting the partitions fail. This was tested with Mageia 3 beta 2. You can overcome this by going to console and doing "modprobe btrfs" manually, after which the installation can continue. Not tested with beta 3 since it came out 4 hours I installed my machine :)

CC: (none) => lihamakaroonilaatikko

i did expert mode, right now, using mga cauldron via netbook, (should be somewhat equal to beta 3), and i had no problem making it btrfs, and continueing, formatting, ... etc...


which iso did you choose, what exact partition combination how many drives did you do, was it a virtual machine if so which type?, etc...


iow: we need an exact reproducer.

Desktop, amd64+2Gb memory, 60G Kingston SSD, no other drives, NVIDIA display card

Bare metal
Boot from USB stick with Beta 2 (made with dd if⦠ofâ¦)
Advanced partition scheme (MBR)
Automatic partition layout
Separate /usr
Change /usr and /home to btrfs
Continue
Say format to all partitions
Continue installation
Results in message: "Unable to mount UUID⦠/mnt/usr"

Workaround:
Go to console
modprobe btrfs
Redo partition
Redo format all partitions
Result: installation continues normally

what happens if you go to custom layout and choose btrfs there?

well debugging a beta2 is now kind of useless, try a netinstall or wait for classical isos.

Looks like it is fixed in Beta3, at least in custom layout path.

(In reply to Thierry Vignaud from comment #34)
> It should already be available providing you switch to expert mode.
> Can you check that?
> I did quite a lot of fixes for btrfs support but I don't remember all the
> details.
> We could enable it in std mode if fsck works fine (Thomas?).


It should work ok, so I think we should show btrfs in standard installer mode for mga3

Fixed in SVN

Status: REOPENED => RESOLVED
Resolution: (none) => FIXED