Carlos, thank you. I ran some tests last night - used 'stress' on hosts to generate high CPU and Memory load. Impressive that ONE almost immediately detects real CPU/MEM usage on the host and won't deploy any VMs to the host. This is what I was looking for and it might work...
Thanks. ________________________________ From: Carlos Martín Sánchez [cmar...@opennebula.org] Sent: Friday, August 02, 2013 5:24 AM To: Dmitri Chebotarov Cc: users@lists.opennebula.org Subject: Re: [one-users] Sharing compute nodes with other applications (i.e. HPC) Hi, The monitorization reports the total cpu and memory in the host, and then opennebula assumes all of it is available for VMs. You can make static adjustments in the scheduler configuration [1], or you could look into the monitorization scripts and try to modify them to suit your needs [2]. A more simple approach would be to disable the hosts [3] when you plan to use them for other jobs. Regards [1] http://opennebula.org/documentation:rel4.2:schg#configuration [2] http://opennebula.org/documentation:rel4.2:img [3] http://opennebula.org/documentation:rel4.2:host_guide#enable_disable_and_flush -- Join us at OpenNebulaConf2013<http://opennebulaconf.com> in Berlin, 24-26 September, 2013 -- Carlos Martín, MSc Project Engineer OpenNebula - The Open-source Solution for Data Center Virtualization www.OpenNebula.org<http://www.OpenNebula.org> | cmar...@opennebula.org<mailto:cmar...@opennebula.org> | @OpenNebula<http://twitter.com/opennebula><mailto:cmar...@opennebula.org> On Thu, Aug 1, 2013 at 5:59 PM, Dmitri Chebotarov <dcheb...@gmu.edu<mailto:dcheb...@gmu.edu>> wrote: Hi I've noticed that ONED continuously monitors compute nodes for available resources (CPU/MEM). So I had this idea of sharing compute nodes between HPC cluster and OpenNebula. The HPC cluster is not necessary loaded 100% all the time and may have spare resources to host VMs. If I added those nodes to ONE, do think sharing resources would work? The idea is when HPC assigns jobs to compute/ONE node, scheduler will monitor the node and "see" that it doesn't have CPU/MEM resource available and won't use it for new VMs.... Thanks. _______________________________________________ Users mailing list Users@lists.opennebula.org<mailto:Users@lists.opennebula.org> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
_______________________________________________ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
