[one-users] scheduler interval

[Hamada, Ondrej]

Hi,
Deployment of VM takes some resources - at least you must load/copy the base image and boot the machine. Both of these operations consumes resources. This might not seem as a problem for deployment of single VM. But deploying tens or hundreds of VMs at single moment can kill your hypervisors/underlying storage (worst case) or at least slow down running VMs - what might kill some apps running in them/against them. So yes, high load really matters.

Ondra

From: users-bounces at lists.opennebula.org [mailto:users-bounces at lists.opennebula.org] On Behalf Of Andrea Gardiman
Sent: Friday, March 07, 2014 4:30 PM
To: users at lists.opennebula.org
Subject: Re: [one-users] scheduler interval


Ok, but I don't understand why it works like a pull model instead of push. Is so because of a too high load matter or only because the initial architectural decision was to work like that and now it is still so?

What would be the drawbacks if were the core,immediately every time a request arrive, tell to the scheduler to schedule that request?

My question comes because I'm working on improving the deployment time of VMs for my master thesis, and I understood that I can choose to deploy VMs little by little, or immediately as they come but I can't see the cases in which deploy VMs little by little would be better that process them immediately.

Thanks a lot,

 Andrea Gardiman.
On Mar 7, 2014 10:58 AM, "Carlos Martín Sánchez" <cmartin at opennebula.org<mailto:cmartin at opennebula.org>> wrote:
Hi,

On Thu, Mar 6, 2014 at 1:30 PM, Andrea Gardiman <andreagardiman at gmail.com<mailto:andreagardiman at gmail.com>> wrote:
Dear all,
I do not understand the usefulness of the scheduler parameter "SCHED_INTERVAL". Why the scheduler act as a periodic task and have to wait a fixed period to schedule all the pending jobs instead of to be event-driven and schedule immediately a request?
What are the benefits and the drawbacks, besides the wait?

Thanks a lot,
 Andrea Gardiman.

The scheduler is a separate daemon. So it polls the core periodically to look for pending VMs.

The usefulness of the SCHED_INTERVAL is that, combined with the other config parameters [1] MAX_DISPATCH and MAX_HOST, allows you to configure a "buffer" to deploy VMs little by little, or immediately as they come.

Regards

[1] http://docs.opennebula.org/stable/administration/references/schg.html

--
Carlos Martín, MSc
Project Engineer
OpenNebula - Flexible Enterprise Cloud Made Simple
www.OpenNebula.org<http://www.OpenNebula.org> | cmartin at opennebula.org<mailto:cmartin at opennebula.org> | @OpenNebula<http://twitter.com/opennebula>


________________________________
This e-mail and any attachment is for authorised use by the intended recipient(s) only. It may contain proprietary material, confidential information and/or be subject to legal privilege. It should not be copied, disclosed to, retained or used by, any other party. If you are not an intended recipient then please promptly delete this e-mail and any attachment and all copies and inform the sender. Thank you for understanding.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20140311/d1e5c6f6/attachment-0002.htm>