2011/8/11 Les Mikesell <[email protected]>
> On 8/11/2011 9:41 AM, Jeffrey J. Kosowsky wrote:
> >
> > > It's basically the same process described in the wiki without:
> > > 1) pre-copying the pool (it would take ages, one of the server only
> has a
> > > 100Mbps internet connection)
> >
> > I'm confused, copying the pool is too slow yet you plan to "plain tar"
> > the pc directory which may be 10's to 100's of times larger (do to
> > pool deduplication)?
>
> He can do this on a per-host basis under pc. Doing each separately is
> likely to work in a reasonable amount of time with rsync (with -H) or
> tar. It won't make the links to pool between different hosts but that
> shouldn't matter on a de-duped zfs.
>
That's what I planned. I can copy a couple of hosts per day, it takes time
but not much attention. It's even scriptable.
Now as Jeffrey said, plain tar will use a lot more bandwidth than
BackupPC_tarPCCopy once the pool is copied. But what about the time it takes
to generate the archive ?
I mean, if BackupPC_tarPCCopy creates its archive as fast as plain tar
transfers the data, it's so much better (maybe less write I/O on the new
server?).
I think I'll have to try some "tar > /dev/null" and "BackupPC_tarPCCopy >
/dev/null" to exclude bandwidth from the equation.
Thanks for your answers :)
Fabrice
------------------------------------------------------------------------------
Get a FREE DOWNLOAD! and learn more about uberSVN rich system,
user administration capabilities and model configuration. Take
the hassle out of deploying and managing Subversion and the
tools developers use with it.
http://p.sf.net/sfu/wandisco-dev2dev
_______________________________________________
BackupPC-users mailing list
[email protected]
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
