[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[rdiff-backup-users] Suggested fs for placing rdiff-backup data stores
From: |
Alex Samad |
Subject: |
[rdiff-backup-users] Suggested fs for placing rdiff-backup data stores |
Date: |
Thu, 11 Feb 2010 11:59:43 +1100 |
User-agent: |
Mutt/1.5.20 (2009-06-14) |
Hi
thought I would break this off the original thread
just to give you some example info on fusecompress datastore
max:/backups/nas/system/boot# du --apparent-size -s --si *
1.3M System.map-2.6.26-1-amd64
1.6M System.map-2.6.30-2-amd64
1.6M System.map-2.6.31-1-amd64
86k config-2.6.26-1-amd64
99k config-2.6.30-2-amd64
102k config-2.6.31-1-amd64
4.0M grub
8.4M initrd.img-2.6.26-1-amd64
7.8M initrd.img-2.6.26-1-amd64.bak
9.8M initrd.img-2.6.30-2-amd64
9.8M initrd.img-2.6.30-2-amd64.bak
11M initrd.img-2.6.31-1-amd64
11M initrd.img-2.6.31-1-amd64.bak
1.8M vmlinuz-2.6.26-1-amd64
2.3M vmlinuz-2.6.30-2-amd64
2.5M vmlinuz-2.6.31-1-amd64
max:/backups/nas/system/boot# du -s --si *
238k System.map-2.6.26-1-amd64
291k System.map-2.6.30-2-amd64
308k System.map-2.6.31-1-amd64
25k config-2.6.26-1-amd64
25k config-2.6.30-2-amd64
25k config-2.6.31-1-amd64
1.8M grub
8.4M initrd.img-2.6.26-1-amd64
7.8M initrd.img-2.6.26-1-amd64.bak
9.8M initrd.img-2.6.30-2-amd64
9.8M initrd.img-2.6.30-2-amd64.bak
11M initrd.img-2.6.31-1-amd64
11M initrd.img-2.6.31-1-amd64.bak
1.8M vmlinuz-2.6.26-1-amd64
2.3M vmlinuz-2.6.30-2-amd64
2.5M vmlinuz-2.6.31-1-amd64
so my primary partition is /backups/.nas and then fuse mounts to /backups/nas/
this is looking at /boot directory for an example. Overall using du I have 1.3
compressed compared to 3.1 uncompressed
du --apparent-size -s --si .nas nas
1.3G .nas
3.1G nas
du -s --si .nas nas
1.6G .nas
1.6G nas
Alex
On Thu, Feb 11, 2010 at 07:45:25AM +1100, Alex Samad wrote:
> On Wed, Feb 10, 2010 at 03:05:14PM -0500, Matthew Miller wrote:
> > On Thu, Feb 11, 2010 at 06:44:46AM +1100, Alex Samad wrote:
> > > > I'm experimenting with LessFS. It's another fuse-based filesystem, and
> > > > in
> > > > addition to compressing blocks, it checksums each block and only stores
> > > > identical blocks once -- "de-duplication". This seems like a particular
> > > > win
> > > > with rdiff-backup, because of the problem with handling of renamed
> > > > files.
> > > thats nice.... what compression tec does it use
> >
> > Read about it yourself here: http://www.lessfs.com/wordpress/?page_id=50
> >
> > In short, it uses a 192-bit hash function (happens to be Tiger) to uniquely
> > identify each block, and then compresses each block with LZO or QUICKLZ.
>
> had a quick read of the web site, just wondering how effective it would
> be with something like rdiff-backup - my line of thinking is that rd
> stores the differences, so I would guess all the original files would
> benefit, but the differences wouldn't
>
> Also with fusecompress you can specify by mime type which files pass
> through ie don't get affected by fusecompress.
>
> I will have to investigate a bit more, run some tests
> >
>
signature.asc
Description: Digital signature
- [rdiff-backup-users] Suggested fs for placing rdiff-backup data stores,
Alex Samad <=