May 18, 20179 yr Hey, Anyone know how to wake up a VM (in my case running Windows 10) from sleep mode, using mouse click or pressing any keyboard's key ?? Thanks!
May 19, 20179 yr Author I am looking for the one the DOES know. for me, using sleep mode is needed. I isolated 6 out of my 8 cores and assigned them to my VM, therefore, when using sleep mode for the VM, 6 cores are NOT being used and their usage drops to 0% until I wake my VM.
May 19, 20179 yr Also, here is a screenshot of one of my servers with 3 vm's currently running on threads 6-23, with threads 0-5 isolated for only unraid/dockers. Sleep is disabled on all vm's.
May 19, 20179 yr Without knowing the specifics of your setup/vm's: IF you really wanted to give it a go using sleep, you would probably need to pass through an entire usb card or port to the vm (for hot plugging.) That "should" allow you the possibility to allow the vm to wake. But it may also trigger the KVM hibernate/sleep state, which I don't believe allows user keystrokes to wake it up. If you're using a physical monitor, you can use monitor sleep without issue and still wake the monitor via keystroke using usb card passthrough (and maybe port passthrough.) When I first started on unRaid about a year ago, sleep caused me multiple problems with vm's until ultimately just disabling it. Now, no problems.
May 19, 20179 yr Author thanks for the replies. well, I am passing thru nVidia GPU and even when I ain't using the VM, the CPU load never drops to 0%... even after disabling the hibernation, sleep, indexing..
May 21, 20179 yr Author On 5/19/2017 at 4:14 PM, 1812 said: I suppose "your mileage may vary." so, tell me the secret of having VM's that doesn't uses the cpu % while they are inactive. does all of your 3 vm's using VNC or GPU ? Edited May 21, 20179 yr by amstel
May 22, 20179 yr 7 hours ago, amstel said: so, tell me the secret of having VM's that doesn't uses the cpu % while they are inactive. does all of your 3 vm's using VNC or GPU ? 6 threads only used by unraid. Rest for vm's. all 3 vm's use an emulator pin set to use one of the 6 unraid threads. 2 use vnc, 1 uses gpu. If you leave vnc connected, it will never idle at next to nothing because the vnc needs CPU power to process. I set each of the vm's to allow thier "monitor" to sleep and disable screen saver (otherwise it is using CPU to run a screensaver on a closed vnc connection- a waste.
May 22, 20179 yr Author 10 hours ago, 1812 said: 6 threads only used by unraid. Rest for vm's. all 3 vm's use an emulator pin set to use one of the 6 unraid threads. 2 use vnc, 1 uses gpu. If you leave vnc connected, it will never idle at next to nothing because the vnc needs CPU power to process. I set each of the vm's to allow thier "monitor" to sleep and disable screen saver (otherwise it is using CPU to run a screensaver on a closed vnc connection- a waste. well, mmy VM configured like this: <cputune> <vcpupin vcpu='0' cpuset='1'/> <vcpupin vcpu='1' cpuset='2'/> <vcpupin vcpu='2' cpuset='3'/> <vcpupin vcpu='3' cpuset='5'/> <vcpupin vcpu='4' cpuset='6'/> <vcpupin vcpu='5' cpuset='7'/> <emulatorpin cpuset='0,4'/> </cputune> screen saver is not configured, I only set the turn off screen after 5 minutes of inactivity. btw, I only use GPU passing and not VNC.
May 22, 20179 yr You only need 1 emulator pin, so use 4. The word on the street is that unRaid prefers to use 0. You also have to realize the core count comes into play when looking at my "0" percent. There is 3% load on one thread, but that rounds down to zero because of the total load. Here is another server running 2 vm's (one OS X, one pfSense-firewall/router for the network) and the following dockers: DuckDNS, Plex, Krusader, NetData. Threads 0-3 are isolated for unRaid and dockers. OS X is on threads 6-9, and pfSense is 4-5. Both vm's use emulator pin 1. Total usage shows 4%. If I were to minimize that down to your 8 cores, then my total usage would be 12% with what is currently active on my server. Take away the pfSense vm and that number would probably be cut in half with your core count. I wouldn't expect for you to hit a consistent 0 utilization since you still have to have some resources used to manage the vm, even when it is doing nothing but running the OS clock and other background tasks. And that number will only get bigger with the fewer cores your have. What are you seeing for overall utilization when the vm is twiddling its idle?
Archived
This topic is now archived and is closed to further replies.