[one-users] Load Balancing ONE controller

Jaime Melis jmelis at opennebula.org
Fri Aug 2 11:21:13 PDT 2013


Hi Simon,

If I understand correctly, by "oned won't scale" you mean that it's not
currently possible to deploy multiple concurrent oned daemons and *not*
that OpenNebula isn't able to handle very large deployments (> 10.000 vms).

In my reply to Dmitri what I meant is that, once we provide a patch to
improve the xmlrpc-server performance, there's no reason, performance-wise,
to run multiple oned daemons. We've seen very large deployments, managed by
a single oned, and the only bottleneck we've seen so far is the
xmlrpc-server. Regarding OpenNebula and its use of the DB and the cache,
that's in fact one of the things that makes the core very fast.


On Fri, Aug 2, 2013 at 12:58 PM, Simon Boulet <simon at nostalgeek.com> wrote:

>
> On Fri, Aug 2, 2013 at 12:21 PM, Jaime Melis <jmelis at opennebula.org>wrote:
>
>>
>>
>> Could you share your scalability experiences with us? Can you a be a bit
>> more specific? What issues have you seen?
>>
>>
> I don't have any scalability issues for the moment.  My backend / oned
> server is dedicate to oned (low-range Core i3 with 4GB RAM). It also runs
> the MySQL database. It's fairly optimized and works great at the moment.
> I'm not using Sunstone, I have a custom front-end that talks to the XML-RPC
> and does some caching (ex. caching templates, VM information for short
> period, etc. ref to # 1967 regarding caching). But I know oned won't scale,
> especially due to the database caching issue. Having an option to disable
> the internal caching opens the door to having multiple oned sharing the
> same DB (which could be a MySQL cluster, or even a NoSQL such as MongoDB if
> someone wanted to write a DB driver for it)
>
> Simon
>



-- 
Join us at OpenNebulaConf2013 <http://opennebulaconf.com/> in Berlin, 24-26
September, 2013
--
Jaime Melis
Project Engineer
OpenNebula - The Open Source Toolkit for Cloud Computing
www.OpenNebula.org | jmelis at opennebula.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20130802/55613015/attachment-0002.htm>


More information about the Users mailing list