Btrfs current status and_future_prospects

Copyright 2014 FUJITSU LIMITED
BtrfsCurrent Status and Future Prospects
Oct 13 2014
Satoru Takeuchi <takeuchi_satoru@jp.fujitsu.com>
Fujitsu LTD.

Background
Core Features
Developments Statistics
Future Prospects
Agenda
1

Background
Core Features
Future Prospects
Agenda
2

Fujitsu has developed Btrfsfor Mission Critical (MC) systems since 2010
Requirements of MC systems
High robustness
•Don’t crash: data duplication
•Error detection: checksum
•Repair, recovery: snapshot, backup/restore, repairing tools
High availability: Should work 365days/24h
•Limited maintenance time: enlarge storage size and backup online
Btrfsis designed for such the requirements
Background
3

Background
Core Features
Future Prospects
Agenda
4

Multi-volumes
Copy-on-Write Style Update
Data/Metadata Checksum
Subvolume
Snapshot
Transparent Compression
Core Features
5

Multi-volumes
Btrfs file system can consists of multiple volumes
Low layered and low overhead than LVM
Many features: RAID, online {add/remove/replace} devices
VFS
File system
(XFS, ext4and so on)
LVM
Block device
VFS
Block device
Btrfs
XFS or ext4 + LVM
Btrfs
# mkfs.btrfs/dev/sd{a,b,c}1
6

Copy-on-Write(CoW) style update
Btrfs uses CoW style data/metadata update
Safer than overwrite style update by design
Overwrite style: Update the data in place
CoW style: Copy, update, and replace pointer
file
data
file
data
file
data
file
data
file
data
data
file
data
data
file
data
7
Will be deleted later
System crash => data keep consistency
System crash => data become inconsistent

CoW versus Overwrite
1,000 surprising power failure test
LinuxFileSystemAnalysisforIVIsystem,MitsuharuIto,Fujitsu
http://events.linuxfoundation.jp/sites/events/files/slides/linux_file_system_analysis_for_IVI_systems.pdf
Result
Ext4: Metadata was corrupted
Btrfs:Workedfinewithoutanyproblem
Inmyinternalsimilartesting,XFScorruptedtoo.
Copyright 8 2014 FUJITSU LIMITED

Btrfs has checksum for each data/metadata extent to detect and repair the broken data
When Btrfs reads a broken extent, it detects checksum inconsistency
With mirroring: RAID1/RAID10
•Read a correct copy
•Repair a broken extent with a correct copy
Without mirroring
•Dispose a broken extent and return EIO
With “btrfs scrub”, Btrfs traverses all extents and fix incorrect ones
Online background job
Data/Metadata Checksum
9

Subvolume
A subvolumeis a file system inside file system
Can be treated as a file system root
•Mountable: most mount options are shared
•Own inodenamespace and quota limit
Efficient: Available space is shared
/
sub
10
# btrfs subvolumecreate sub

Snapshot
Copy of a subvolume
Far faster than LVM
•Not a full copy, but only update metadata in CoW style
Readonly snapshot: with –r option
Incremental snapshot: snapshot of snapshot
snap
sub
2
2
2
2
2
2
# btrfs subvolumesnapshot [-r] ./sub ./snap
A
B
C
11
sub
1
1
1
1
1
1
A
B
C
Reference count
snap
sub
1
1
2
1
2
2
A
B
C
C’
1
1
1
Capture a snapshot
Update data C in a snapshot

Performance of Snapshot: Btrfs versus LVM
1.Copy the following data to a volume
Consists of 100 directories and 100 files for each directory
•File size: 1MB
2.Capture a snapshot of the volume
Hardware Environment
Software Environment
•PRIMERGY RX300 S6
•CPU:Intel Xeon X5690 3.47GHz x12 core
•Memory:16GiB
•Storages:100GB HDD x 2
•Red Hat Enterprise Linux 7.0
•Filesystems
•Btrfs
•Data/metadata: RAID1
•Other options: default
•XFS:default options
•Volumemanager for XFS
•dm-thinp:chunksizeis 256KiB
•LVM:RAID1

Result
Copy: Btrfs > LVM >>> dm-thinp
Snapshot: Btrfs > dm-thinp >>> LVM
Volumetype
Copy
Snapshot
Without page cache
With page cache
Btrfs
106s
0.126s
11.7s
XFS on dm-thinp
209s
0.260s
15.5s
XFSon LVM
133s
1.03s
45.2s

Transparent compression
Automatically compress/expand file data on I/O
Low space consumption and high I/O performance
•Need some extra CPU time
Usage: mount -o compress={lzo,zlib} <device><mntpoint>
•Can also be enabled/disabled for each file
compress/expand
system
storage
Page cache
without compression
with compression

Background
Core Features
Developments statistics
Future Prospects
Agenda
15

Patch statistics
Performance
Summary
Developments statistics
16

Patch Statistics
17 Copyright 2014 FUJITSU LIMITED

Patch Statistics: Tips of v3.17
Rejected by Linus

Patch statistics: Main changes
Inodeproperties
offline dedup
Improve sync write ~60%
RAID5/6
replace subcommand
quota
send/receive
btrfsck
repair
auto defrag
scrub
Improve error handling
Inodeproperties

Fujitsu’s contribution
•btrfsck, error handling
•fast {random/async} write
•LZO compression
•read only snapshot
•random Bug fixes
•enrich xfstests

Fujitsu’s contribution: btrfs-progs
21
•fsck
•error handling
•random bug fixes
•enrich xfstests
•documentation

Performance measurement
Hardware Environment
Software Environment
•PRIMERGY TX300 S6
•CPU: Xeon x5670 x 2
•12 core
•HT is disabled
•Memory: 4GB
•HDD: 300GB x 1
•MegaRAIDSAS, HITACHI HUS156030VLS600
•Benchmark software: filebench
•Kernel: 3.14.11, 3.15.4, 3.16.3, and 3.17-rc2
•I/O scheduler: deadline
•File systems: Btrfs(single volume), XFS, and ext4
•default mkfsoptions and mount options

The result: Compare with other file systems
23
Kernel version: v3.17-rc2

The result: Compare with old Btrfses

VFS has also improved performance
Accomplished by VFS layer performance enhancement

Summary
Ready to use without RAID5/6
Performance: OK
Stability: OK
•# of new features has decreased
•Test coverage has increased
Features: almost OK
•RAID5/6: Lack of scrub and replace subcommands
RAID1 and RAID10 are the best choice
Especially safe and stable

Background
Core Features
Developments statistics
Future Prospects
Agenda
27

Future Prospects: Fujitsu’s plan
RAID 5/6 enhancement
Add scrub and replace subcommands
•We’re testing patches now and will post it to linux-btrfsML soon
Add five tests for these features to xfstests
Further enhancement of robustness and performance
Repairing tools and so on
Education and documents for this purpose
Operation know-how
•Btrfsoperations are different from other file systems
•e.g.Btrfsの基礎part1機能編(It’sinJapanese.NowtranslatingtoEnglish…)
http://www.slideshare.net/fj_staoru_takeuchi/btrfs-part1
File system structure
Code logic

Future Prospects: Btrfs users are increasing
Will be used by OpenSuSE13.2 as its default
Supported by Ubuntu
Available with RHEL7 as tech-preview
Will be used for In Vehicle Infortaiment(IVI) system

Conclusion
PleasetryBtrfs
It’sreadytouse
RAID1/10arethebestchoice
RAID5/6needsomemorework
Recommendtheneweststablekernel

References
Btrfsの基礎part1機能編
http://www.slideshare.net/fj_staoru_takeuchi/btrfs-part1
Linux-btrfsMLlinux-btrfs@vger.kernel.org
Btrfswiki
https://btrfs.wiki.kernel.org/index.php/Main_Page

Btrfs current status and_future_prospects

More Related Content

Btrfs current status and_future_prospects