[rear-users] Article about ReaR in German Linux Magazin

oh oliver at g.dom.de
Mon Mar 14 02:56:42 CET 2016

Hi Schlomo,

thank you for this clarification. Now mkrescue and mkbackup are 
perfectly clear :)

Just two more things I was wondering about while testing ReaR on a USB 
stick. If I use NETFS is there a way to cut out files/folders? Something 
like the exclude option of rsync.

I used rear mkbackup on quite a new laptop. It has a 128GB SSD running 
debian stretch. With that I booted the rescue system on a rather old 
laptop with 80 GB HD. Everything should fit as the source system has 
less than 16 GB in total. Problem is the partitioning of the disk. It 
stops with: parted invalid token: logical. I changed things in 
layout/diskrescue.sh but to no avail.
I had a look with parted and found two primary partitions which made no 
sense. I would expect that the script would suggest something useful and 
ask for approval. This is not really an issue. I was just playing around 
but I could provide you with more details.



> ​Hi Oli,​
> On 11 March 2016 at 17:40, Oliver Hoffmann <oliver at g.dom.de 
> <mailto:oliver at g.dom.de>> wrote:
>     something I need to get straight about mkrescue. If I just do a rear
>     format and mkrescue creating a USB stick and boot it up then I
>     would be
>     dropped to the rescue shell.
> ​I would call it "boot the rescue system", but yes.
>     If I just do a rear rescue then the result would be a naked system
>     as it
>     was before but without any data or even extra binaries, won't it?
> You still have to run rear recover to see any action. ​Probably doing 
> that exactly as you wrote will actually wipe your system, reformat the 
> disks and prompt you to restore the files into /mnt/local.
>     Means this is only useful if I try to repair the system rather than
>     carry out a rescue.
> ​Yes, you could use the rescue system to repair a system. However, 
> that is not the main purpose.
>     In other words in most of the cases mkbackup is what one wants,
>     isn't it?
> ​That depends on the value of the BACKUP variable in your config (see 
> rear dump​).
> If you use one of the internal backup methods like BACKUP=NETFS then 
> you are correct and users should use rear mkbackup. The difference 
> between rear mkrescue and rear mkbackup is that rear mkrescue will 
> only create the rescue image while rear mkbackup will both create the 
> rescue image and also create a new backup.
> ​If you use an external backup method, e.g. BACKUP=TSM, then rear 
> mkbackup and rear mkrescue​ do exactly the same: Create a new rescue 
> image. In this mode the assumption is that your external backup system 
> already has a full (or relevant) backup of your system. Running rear 
> recover in the rescue system will then wipe the system, recreate the 
> partitions and file systems and instrument your external backup 
> software (e.g. TSM) to actually restore the files.
> To understand this distinction you need to see the original purpose 
> for which we created ReaR: Doing bare metal disaster recovery with 
> backup software that does not support this. Or where the bare metal 
> addon is very expensive.
> I wrote rear 1.0 in 10 days for a client who had Galaxy as his backup 
> software. The bare metal addon of that backup software was sold at 
> about 1000€ per server. With the amount of servers that the client 
> wanted to cover (about 50), contracting me to write rear was actually 
> much cheaper for them :-) And rear worked much better than the 
> proprietary bare metal solution because it supported all custom 
> drivers and worked fully automated.
> HTH,
> Schlomo
> _______________________________________________
> rear-users mailing list
> rear-users at lists.relax-and-recover.org
> http://pikachu.3ti.be/mailman/listinfo/rear-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pikachu.3ti.be/pipermail/rear-users/attachments/20160314/cbea2d07/attachment.html>

More information about the rear-users mailing list