[pve-devel] features ideas for 3.0 roadmap

Alexandre DERUMIER aderumier at odiso.com
Wed Mar 13 09:39:09 CET 2013


>>Why do you thing cpu usage is relevant? 
>>I guess server load, memory and disk usage are more important. 

What I have in my mind was live migrate vm if a server is overloaded with cpu.

- For memory we already have ksm && ballooning locally, 
- For disk, as we live migrate, it must be a shared storage, so migrate the vm will not help.

But Indeed we could also check memory or network link for example.

Vmware seem to call it "VMware Distributed Resource Scheduler (DRS)", I never use it, but here a demo:
http://www.youtube.com/watch?v=aC1F_PMslmg

They seem to handle cpu && memory overload.



----- Mail original ----- 

De: "Dietmar Maurer" <dietmar at proxmox.com> 
À: "Alexandre DERUMIER" <aderumier at odiso.com>, "James Devine" <fxmulder at gmail.com> 
Cc: pve-devel at pve.proxmox.com 
Envoyé: Mercredi 13 Mars 2013 07:34:52 
Objet: RE: [pve-devel] features ideas for 3.0 roadmap 

> - if host total cpu > 80% since X minutes 

Why do you thing cpu usage is relevant? 
I guess server load, memory and disk usage are more important. 


More information about the pve-devel mailing list