On 09/29/2014 04:50 PM, Jeff Schroeder wrote:
At my last employer, we used cobbler with thousands of nodes. In fact,
due to it being unbearably slow due to yaml, we convinced Michael to
switch to JSON as the preferred serialization format. We had the
setting on to prevent duplicate ip addresses as we used static dhcp
reservations for everything and a pxe bootable infrastructure (custom
operating system that ran HPC like jobs with the os in memory).
I've seen several places using cobbler to scale to 20+ master servers
and ultimately thousands of system records. Just because you don't use
it doesn't mean it doesn't exist. I mostly lurk as cobbler solves my
needs extremely well, but that comment needed a reply.
Thanks for the feedback, Jeff. Indeed most people I talk to use Cobbler
with few systems
and I incorrectly generalized.
It amazes me that one Cobbler server can handle hundreds of systems
storing its objects in the
file system and loading all of them to RAM at Cobbler startup. I'd
expect a database backend and a
on-demand load for that degree of scalability.
Regards,
Alan Evangelista
_______________________________________________
cobbler-devel mailing list
cobbler-devel@lists.fedorahosted.org
https://lists.fedorahosted.org/mailman/listinfo/cobbler-devel
