You are not logged in.

#1 2009-02-13 09:32:50

Llama
Banned
From: St.-Petersburg, Russia
Registered: 2008-03-03
Posts: 1,379

btrfs and Arch

Hi,

As of kernel 2.6.29 there's going to be support for btrfs. From what I hear, its already usable, albeit at one's own risk smile . Anybody experimenting here (the forum search turned out essentially nothing)?

Offline

#2 2009-02-13 12:06:51

Ranguvar
Member
Registered: 2008-08-12
Posts: 2,577

Re: btrfs and Arch

Offline

#3 2009-02-13 12:08:43

iBertus
Member
From: Greenville, NC
Registered: 2004-11-04
Posts: 2,228

Re: btrfs and Arch

Is it just me or does that thread only have 2 pages currently?

Offline

#4 2009-02-13 12:13:35

proxima_centauri
Member
From: Nova Scotia, Canada
Registered: 2008-07-17
Posts: 117
Website

Re: btrfs and Arch

I see 3 pages.


Thinkpad T61p - 15.4' WSXGA TFT - 2.5Ghz Intel Core2 Duo T9300 - 2X2GB Kingston RAM - 160GB 7200RPM - NVIDIA Quadro FX 570m - Intel 4965AGN

Offline

#5 2009-02-13 12:42:55

RaisedFist
Member
From: Romania
Registered: 2007-01-30
Posts: 556
Website

Re: btrfs and Arch

depends on your settings

Offline

#6 2009-02-13 12:53:08

Nezmer
Member
Registered: 2008-10-24
Posts: 559
Website

Re: btrfs and Arch

You should know that BTRFS is still in early stage of development . The disk format is not finalized yet and new features are still getting implemented . Every time the disk format changes you will have to reformat .


English is not my native language .

Offline

#7 2009-02-13 13:12:16

jelly
Administrator
From: /dev/null
Registered: 2008-06-10
Posts: 716

Re: btrfs and Arch

why don't u use BTRFS in a virtual machine ?;)

Offline

#8 2009-02-13 13:28:53

Zariel
Member
Registered: 2008-10-07
Posts: 446

Re: btrfs and Arch

AFAIK Btrfs disk format will be kept pretty stable now its in the mainline kernel, and it was hoped to be released Q408, which may still happen but i doubt it.

Offline

#9 2009-02-13 15:44:24

Ranguvar
Member
Registered: 2008-08-12
Posts: 2,577

Re: btrfs and Arch

btrfs will be, by all accounts, in 2.6.29 (unless obviously if something goes very wrong).

From what I understand, the format is either done or close to it, and most features are done, but it's still very slow as it has not yet been optimized. This is all hearsay, though.

Offline

#10 2009-02-13 17:04:39

sehnpaa
Member
Registered: 2008-03-29
Posts: 3

Re: btrfs and Arch

This is what main developer Chris Mason wrote when Btrfs 0.17 was released a month ago:

"In general, the disk format isn't going to change unless we find a
critical bug with the old format. Every attempt will be made to
maintain backwards compatibility."

http://www.mail-archive.com/linux-btrfs … 01645.html

AFAIK, the disk format haven't changed since then.

Offline

#11 2009-02-13 17:35:02

lucke
Member
From: Poland
Registered: 2004-11-30
Posts: 4,019

Re: btrfs and Arch

Wanted to compare unpacking time on ext4 and btrfs, here's the btrfs one, khkh:

[18:32:23][root@kingdom]# time tar xf openoffice-base-3.0.1-1-i686.pkg.tar.gz
zsh: segmentation fault  tar xf openoffice-base-3.0.1-1-i686.pkg.tar.gz
Real: 9.12s User: 0.07s System: 1.92s Percent: 21% Cmd: tar xf openoffice-base-3.0.1-1-i686.pkg.tar.gz

(it seems the partition got full, but well, it didn't handle it gracefully)

Last edited by lucke (2009-02-13 17:37:15)

Offline

#12 2009-02-13 17:58:01

lucke
Member
From: Poland
Registered: 2004-11-30
Posts: 4,019

Re: btrfs and Arch

Okay, very rigid tests (with standard mkfs/mount options and dropping caches in between tests):

btrfs without compression: 
Real: 4.74s User: 1.30s System: 2.58s Percent: 82% Cmd: tar xf kernel26-2.6.28.5-1-i686.pkg.tar.gz
/dev/sda7             510M  224M  287M  44% /home/lucke/h

btrfs with compression:
Real: 4.53s User: 1.30s System: 2.40s Percent: 81% Cmd: tar xf kernel26-2.6.28.5-1-i686.pkg.tar.gz
/dev/sda7             510M   59M  452M  12% /home/lucke/h

ext4:
Real: 3.21s User: 1.52s System: 0.75s Percent: 70% Cmd: tar xf kernel26-2.6.28.5-1-i686.pkg.tar.gz
/dev/sda7             494M  129M  340M  28% /home/lucke/h

edit: I got 24% space occupied on subsequent runs of btrfs without compression. Human error?
edit2: And it's 14% for the second one. Delayed allocation in this case, I presume.

Last edited by lucke (2009-02-13 18:14:43)

Offline

Board footer

Powered by FluxBB