2. What is ReaR
ReaR is a highly modular backup/recovery framework, used in
case of disaster to restore GNU/Linux operating system on bare
metal machine.
Basically ReaR is tar-gziping/ (slash) + stores metadata
regarding file systems / LVM / boot loader / HW and SW RAID
configuration, etc packed in initrd image along with system
tools – altogether usedso that same FS layout / boot
configuration can be recreated upon restore
Mature product - created as a spin-off of two existing projects:
OpenVPN GW builder (Schapiro) and mkcdrec (D’haese)
Deutsche Qualität
3. Why ReaR
Open source software written in shell using standard UNIX utils -
can easily be modified to suit environment needs
Fast – both backup and restore operation
Online – no service interruption during backup
Space efficient – only effective data is being backed up
Can be easily modified for specific needs and extend with
new features
Can restore to another (similar) hardware – hence system
cloning is possible
Can also be used to manually restore individual files
Can backup / restore to different backend storage (TSM, NFS)
4. Why not ReaR
Does not work for Linux systems older than ~ 10 years (<2.6)
Can work only if systems are on the same broadcast domain
(on PXE based bare metal restore)
Can fail in heavily customized GNU/Linux installation
5. Our modifications
Patched sshd.conf/passwd to permit remote root logins
Disk geometry fix patch
grub.conf fix patch
MAC addressed list also included in backup meta data
DHCP/manual IP address assignment during restore (restore in
different network)
6. Managing backups
There wasn’t free tool for managing backups – so we wrote
our own:
rearadm – ruby scrip (easy to understand – clean code)
~ 1000 lines (including integrated man pages)
archiving, versioning and deleting backups
creating, changing and deleting hosts
provisioning backups for restore procedure
cloning systems