a reference in a database is essentially the same thing as a hardlink on a
filesystem as far as function. the gains from a database would be no inode
creation and depending on the filesystem, faster hardlink and file creation
or deletion. filesystems must be good at a number of things where a
database is really good a a few things.
a filesystem is a more flexible way to store files but a database is much
more efficient at accessing data. in the past, attempts to make a
filesystem out of a database has not been very good because the database
isnt good at a number of things that a filesystem is good at, such as large
files and simplicity in system structure. many modern filesystems borrow a
lot for a database but maintain that simplicity.
i agree that a database would be very good for the backup process as well as
the retrieval process BUT a filesystem is better at mirroring a live system
as well as processing larger files.
there are good arguments both ways here. the current hardlink system is
much more transparent to a user.
i have given up on the mission to get backuppc to use a database, but have
looked into some projects that put a 'virtual' (meaning each attribute of
ext2 is converted to an appropriate sql function) ext2 filesystem in a
mysql database. that would make the database access transparent to backuppc
but it would gain all the speed of not creating inodes and hardlinks as well
as speed running the nighty cleanup. it would unfortunately degrade the
speed of storing large files as a regular filesystem just does that faster.
On Feb 7, 2008 8:19 AM, Nils Breunese (Lemonbit) <[EMAIL PROTECTED]> wrote:
> Stephen Joyce wrote:
>
> > On Wed, 6 Feb 2008, Les Mikesell wrote:
> >
> >> You also have to know how many references there are to each pool
> >> item.
> >> That is, pretty much duplicate the code of a filesystem without
> >> gaining
> >> much. And you can't let any of this change for the duration it
> >> takes to
> >> complete your mirroring.
> >
> > I understand how the hardlinks are used (and it works pretty well)
> > but I
> > can't help but think that a database of file references would work
> > as well
> > (possibly better) than the hardlinks...
>
> I really doubt that a database would perform any better. If it did,
> then I guess we need better filesystems (I nicked this argument from
> someone on this list, it was used recently when someone suggested that
> a database would be a more efficient way of storing files than a file
> system).
>
> Nils Breunese.
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> BackupPC-users mailing list
> BackupPC-users@lists.sourceforge.net
> List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
> Wiki: http://backuppc.wiki.sourceforge.net
> Project: http://backuppc.sourceforge.net/
>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
BackupPC-users mailing list
BackupPC-users@lists.sourceforge.net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
