So, we do this 2 ways in my shop: 1). We have a full replica system in another data center, which, along with having all the system/profile/repo/file/etc data replicated, serves systems in that DC w/ dhcp/pxe/kickstart etc. All changes are made on our primary and a sync trigger kicks off a replication run to the "slave" system. 2). Both systems get their /var/lib/cobbler (where all the configs live) backed up hourly.
Everything else (everything that ends up in /etc/cobbler/, etc) including the install, populating our few hundred subnets in to the dhcp.template, etc etc, is done via puppet (which is in a backed up git repo). So, I can build a new cobbler instance about 99% hands off, with the 1% being restoring the contents of /var/lib/cobbler from backup and running a sync. On Thu, May 8, 2014 at 2:54 PM, Tim Skirvin <[email protected]> wrote: > What is the current best practice for backing up and restoring a > cobbler server's data, as part of a disaster recovery scenario? > > The best I can see in the online documentation is to use the > 'replicate' functionality; but that requires rsync, which seems an odd > requirement (why not allow ssh?), and it also requires that you have a > single working cobbler server in the first place. > > Maybe another way to ask the question: if I wanted to restore my > cobbler configuration from tape, what would I have to do besides restore > /var/lib/cobbler/* and /etc/cobbler/settings from backup and run 'cobbler > sync'? > > - Tim Skirvin ([email protected]) > -- > HPC Systems Administrator / Developer > http://www.linkedin.com/in/tskirvin > USCMS-T1 Collaboration Fermilab Scientific Computing > > _______________________________________________ > cobbler mailing list > [email protected] > https://lists.fedorahosted.org/mailman/listinfo/cobbler > > -- Matthew Nicholson
_______________________________________________ cobbler mailing list [email protected] https://lists.fedorahosted.org/mailman/listinfo/cobbler
