[Rear-users] Hi to the ReaR community

Jeroen Hoekx jeroen.hoekx at hamok.be
Thu Jan 26 08:24:21 CET 2012


On 26 January 2012 00:44, Les Mikesell <lesmikesell at gmail.com> wrote:

> I think it is much easier/safer to hack things when (a) the pieces are
> very compartmentized, and (b) local changes are clearly separated from
> the base system that may change in updates.   I'm very much interested
> in the parts of ReaR that other backup systems don't handle
> (specifically getting the filesystems re-created with any underlying
> raid/lvm to match the original), but I haven't made the effort to
> understand it so far because it seems too intertwined with everything
> else with the real work hidden behind some abstractions that look easy
> to break.  Maybe that's just a misconception but I can't find much in
> the way of design documents and shell scripts aren't all that
> readable.
True. No design docs exist. There's a document docs/layoutman.txt
which shows the general flow of things. The abstractions are just two
files. One to describe the layout (disklayout.conf). Rear generates
code to recreate the system based on that description in another file
(diskrestore.sh), which is executed.

In Rear it's not that hard to understand something. Just run rear -s
mkbackup and all scripts will be shown to you. Need the layout code?
Just look at the ordering of scripts under layout/. Their names along
with the document will get you a long way.

> I'm also interested in how to bring up (multiple) network
> interfaces with the original IP's when the restore goes on a different
> chassis, but again, haven't gotten that far.

I also think we need some work here in Rear because the use cases for
more flexibility are clearly there, both in the DR and cloning
situations.

Greetings,

Jeroen




More information about the rear-users mailing list