Rustdesk intermittently slow accessing proxmox vm manjaro xfce

I have 2 manjaro xfce vm’s running in a Proxmox VE. each have a copy of kmine ( Iknow, why Kmine, it is an app that I can quickly use to judge response time when accessing either of the manjaro vm’s using rustdesk access.)
One Manjaro VM is very responsive all the time. The other, after sitting idle for a while, is very sluggish.
What settings in Manjaro can effect rustdesk access to a VM ?
If i reboot the slow manjaro, rustdesk IS as resposive as the other manjaro.
Any thoughts ?

I’m very picky about smoothness in my flight simulators, running at 1440p@167 Hz. (All in virtual.)

It took me a lot of troubleshooting and trial and error to make it smooth, turns out it was something you have to tweak, that’s unique to Xen 3 architectures or older.

But we have no information about anything. And I don’t know Proxmox all that well.


In the guest, what is the output of: inxi -zv8


Most of the problems are video related to these types of things. But the way you are doing video is making it more unique.

Ruskdesk will attempt to use hardware video encoding from the guest. So most likely you will be doing this in software. (Only some of my VMs can do this with GPU pass-through.) I also don’t use Rustdesk, just waypipe or X11.

You also need guest tools installed for any performance. I know what to do with qemu, virtualbox, vmware, but no idea on proxmox guests.

This includes faster performing device drivers for disk, network, video, etc.

What I’m seeing is acceptable preformance on one Proxmox Manjaro VM and unacceptable performance running a simply app, Kmine. I’m only using kmine as a way to compare 2 VM’s performance. Both VM’s are running Rustdesk, so rustdesk does not appear to be the issue, and both are running in Proxmox, so Proxmox does not seem to be the answer either. Leaving a setting in one Manjaro vm verses the other. I’m trying changing display and system sleep and hybernation setting, so far to no success.
The other odd obervatrion is, if I reboot the failing Manjaro both kmines operate the same. Only one VM slows a LOT if I exit rustdesk from both vm’s and wait overnight or some undetermined time.

That would actually give us an idea of how the guest is setup. I didn’t realise the options are pretty close to qemu.

As far as I know, sleep (or hibernate) is not supported in most hypervisors, including proxmox. They stay on. And when you don’t need them you shut them off, you do not hold on to their memory state. Often there are dozens of VMs running at once on hypervisors.

The virtio guest drivers don’t even support sleep. So it would hang for me as well, if I tried to do it.

Getting fast video in a from a virtual guest is not straight forward. Remote consoles (or remote apps) are generally really slow over the network. Even when that network is on the same computer.

It’s the difference of the CPU literally talking directly to your RAM, but instead it has to jump through a hundred loopholes before it can get to it.

Why one messes up? Sometimes? I have no way to tell with just this information.

There are things you can do to make them much faster. For this Rustdesk to be fast you would essentially need a gaming 3D GPU in the guest. (It doesn’t have to be a gaming GPU, but it needs to be able to encode H.264/HEVC/AV1/etc. Your VMs will not be able to do this by default. Proxmox even supports GPU pass-through, if you have two of them.)

Since you are so set on using kmines try using it through waypipe instead of this Rustdesk. Just to see if it could be causing issues.

Make sure waypipe is installed on both. (It probably is installed by default now?)

Then on your host run it:

waypipe --no-gpu ssh user@host kmines



I played around with “power management” in Manjaro and these setings work so both Manjaro’s provide the same response when running Kmine.

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.