Hi,
I am trying to understand Xen's save VM state works.
I have use case where I am running 3 VM's on a 4-core/8GB server with
1core/2GB per VM. Each of the VM runs an HPC application, and most of the
time 1.5GB memory is allocated to the application running in the VM.
Now I am trying to do is to pause VM-1 and VM-2. Save their respective
states. Reduce their memory and cap CPU to lowest possible value; then to
alloacate the taken resource from VM-1 and VM-2 to VM-3 for X amount of
time. Once the application in VM-3 is finished, then I reduce it's
CPU/Memory to pre-start stage and re-allocate 1CPU/2GB again to VM-1 and
VM-2
My question is; would the application in VM-1 or VM-2 will be able to start
executing from the last saved point. Is there a way in Xen to check-point
this?
Thanks,
Omer
**************************************
CERN, Engineering Department,
CH-1211 Geneva 23, Switzerland
Email: Omer.Khalid@cern.ch
**************************************
I am trying to understand Xen's save VM state works.
I have use case where I am running 3 VM's on a 4-core/8GB server with
1core/2GB per VM. Each of the VM runs an HPC application, and most of the
time 1.5GB memory is allocated to the application running in the VM.
Now I am trying to do is to pause VM-1 and VM-2. Save their respective
states. Reduce their memory and cap CPU to lowest possible value; then to
alloacate the taken resource from VM-1 and VM-2 to VM-3 for X amount of
time. Once the application in VM-3 is finished, then I reduce it's
CPU/Memory to pre-start stage and re-allocate 1CPU/2GB again to VM-1 and
VM-2
My question is; would the application in VM-1 or VM-2 will be able to start
executing from the last saved point. Is there a way in Xen to check-point
this?
Thanks,
Omer
**************************************
CERN, Engineering Department,
CH-1211 Geneva 23, Switzerland
Email: Omer.Khalid@cern.ch
**************************************