Hi,
I am a little confused about how SpotCloud schedules VMs. From SpotCloud Setup Guide, it seems that only aggregate capacities need to be specified (number of concurrent VMs, maximum number of CPUs etc.) Given the fact that there may be multiple hardware profiles, how does SpotCloud decide how many VMs of each hardware profile that can be scheduled, where (on which machine) they are scheduled?
Thanks,
Eric
It sounds to me like what you're describing is "how is each provision
job distributed in each provider's environment?" This is entirely
handled by the provider. All SpotCloud does is says "please create a
new VM using this hardware profile". This will either pass or fail.
All the hardware profile does is say what physical hardware will be
allocated to the VM. If the provider can't fullfill this request, the
job will fail.
Hope this helps clarify things a little.
Adam