Chris O'Regan wrote:
> If I was maintaining a disaster recovery setup, I'd just use
"cobbler
> replicate" between the two servers on cron, or if they are supposed to
> always be in sync (and cron is unacceptable) and there isn't a reliable
> concept of a master, possibly set up cobbler triggers to rsync
> /var/lib/cobbler/config to the other server whenever something changes
> on one of them.
>
Interesting, I did not notice that on the wiki when I was hunting for
information this afternoon.
Thanks for the heads up! I'll put it up there.
Yeah, this sounds like what I want. Is this
a push or a pull?
It's a pull.
I gather that "cobbler replicate" is run on each
subcobbler (pull), right?
Correct. Subcobblers can have subcobblers if you so wish :)
Hrmm...my "ks_mirror" directory on my dev box
is pretty much empty (just a "config" directory). Am I missing
something? What should be in there?
That's the directory used by "cobbler import", if you were sourcing
trees externally (without mirroring them) this data would not be in use.
Is this going to replicate my kickstarts and snippets, too, or do I
have
to find a way of sharing them? If the latter, can those be put onto NFS
safely? What mechanism is used for replication?
Yes, it can help move over as little or as much as you want. See
"cobbler replicate --help"in 1.2.4. If it can't do something you want,
let me know, and we can see about adding it.
Replication is achieved using Cobbler XMLRPC for basic data, and files
are transferred using scp/rsync. Usage of ssh-agent before cobbler
replicate is therefore recommended unless you are using authorized_keys.
Thanks,
Chris
_______________________________________________
cobbler mailing list
cobbler(a)lists.fedorahosted.org
https://fedorahosted.org/mailman/listinfo/cobbler