SPICE performance sluggish

[ml1941]

PVE 6.3-2. Configuring VMs with 8GB RAM, and SPICE qxl, memory=32. Tried several of the latest distributions, specifically Manjaro, ArchLabs, and Fedora Workstation. Performance is just very slow to the point that it is frustrating to use. Verified that spice-vdagent is running and that xf86-video-qxl package was installed (in the case of the Arch-based distros). Switching to VNC yields better performance, but I really want to use SPICE if possible. Has anyone else experienced this? What can I try to debug?

For example, the Manjaro guest is running like:

/usr/bin/kvm -id 104 -name manj1 -no-shutdown -chardev socket,id=qmp,path=/var/run/qemu-server/104.qmp,server,nowait -mon chardev=qmp,mode=control -chardev socket,id=qmp-event,path=/var/run/qmeventd.sock,reconnect=5 -mon chardev=qmp-event,mode=control -pidfile /var/run/qemu-server/104.pid -daemonize -smbios type=1,uuid=4a1538aa-ef64-40bd-a811-e4713d1a2c0f -smp 4,sockets=1,cores=4,maxcpus=4 -nodefaults -boot menu=on,strict=on,reboot-timeout=1000,splash=/usr/share/qemu-server/bootsplash.jpg -vnc unix:/var/run/qemu-server/104.vnc,password -cpu kvm64,enforce,+kvm_pv_eoi,+kvm_pv_unhalt,+lahf_lm,+sep -m 8192 -device pci-bridge,id=pci.1,chassis_nr=1,bus=pci.0,addr=0x1e -device pci-bridge,id=pci.2,chassis_nr=2,bus=pci.0,addr=0x1f -device vmgenid,guid=7a044272-7ce1-4e7a-9932-84f3d7dad62b -device piix3-usb-uhci,id=uhci,bus=pci.0,addr=0x1.0x2 -device qxl-vga,id=vga,vgamem_mb=64,ram_size_mb=256,vram_size_mb=128,max_outputs=4,bus=pci.0,addr=0x2 -chardev socket,path=/var/run/qemu-server/104.qga,server,nowait,id=qga0 -device virtio-serial,id=qga0,bus=pci.0,addr=0x8 -device virtserialport,chardev=qga0,name=org.qemu.guest_agent.0 -device virtio-serial,id=spice,bus=pci.0,addr=0x9 -chardev spicevmc,id=vdagent,name=vdagent -device virtserialport,chardev=vdagent,name=com.redhat.spice.0 -spice tls-port=61000,addr=127.0.0.1,tls-ciphers=HIGH,seamless-migration=on -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 -iscsi initiator-name=iqn.1993-08.org.debian:01:dad8811ab248 -device virtio-scsi-pci,id=scsihw0,bus=pci.0,addr=0x5 -drive file=/dev/pve/vm-104-disk-0,if=none,id=drive-scsi0,cache=writeback,format=raw,aio=threads,detect-zeroes=on -device scsi-hd,bus=scsihw0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0,id=scsi0,bootindex=100 -netdev type=tap,id=net0,ifname=tap104i0,script=/var/lib/qemu-server/pve-bridge,downscript=/var/lib/qemu-server/pve-bridgedown,vhost=on -device virtio-net-pci,mac=46:05:C1:C0:7F:84,netdev=net0,bus=pci.0,addr=0x12,id=net0,bootindex=101 -machine type=pc+pve0

 

[tom]

I assume you run Wayland inside?

=> switch to Xorg.

 

[ml1941]

No, the Arch based distros are both Xorg. The Fedora install is Wayland, but I just installed that for comparison... I really want to run a graphical Arch-based VM, either Manjaro or Archlabs.

 

[tom]

I primarily use Debian Buster or Ubuntu (with Xorg), works great with SPICE. I am not a Arch user, maybe someone else can comment on that.

 

[ml1941]

I will try one or both of those and comment here. I'm interested in trying to isolate where the issue may be.

 

[ml1941]

I will try one or both of those and comment here. I'm interested in trying to isolate where the issue may be.

It is marginally better with Debian 10 guest, default XFCE desktop, but still not usable for constant interactive use. Tried Windows 10 virt-viewer client as well as virt-viewer on a Linux desktop--performance is not materially different between the two clients.

@tom, are Youtube videos smooth for you on your setup?

 

[tom]

@tom, are Youtube videos smooth for you on your setup?

Yes. How fast is your network between client and VM, latency?

 

[ml1941]

Yes. How fast is your network between client and VM, latency?

It is across the internet. Just measured with ping, average 23ms.

 

[ml1941]

OK, after playing with Debian a bit more, I think it is considerably better. Video from youtube still not working great but definitely better than Manjaro.

Are there any quantitative, data-driven measures of graphical remote performance that can be useful for comparison/benchmarking? The only one I can think of right now is glxgears... on Debian the FPS that outputs is about double that from Manjaro.

 

[kiraqian]

I have the same issue with SPICE for manjaro guest. noVNC looks better. My solution is install nomachine (server) on manjaro and install nomachine (client) on the PC I would like to remote to manjaro. Nomachine works very great in this case because you are directly connect to manjaro not through the PVE proxy. But each time after reboot, you have to login to manjaro with SPICE then switch to nomanchine connection. Otherwise the display resolution is low.

 

[mbx]

@tom SPICE + Wayland is still horribly slow in VMs. Are there any plans to address these issues? Xorg is not really what I want.

 

[duaneg]

I know this is old, but just for future reference .. I don't know why or what it has to do with exactly, but I typically have used SPICE viewer without any issue until today.

Today my XFX Radeon RX580 stopped working properly and I was no longer able to display my 4 monitor setup as normal, so I had to switch to a spare NVidia GeForce GTX 1060 6GB. All is well, driver upgrades are completely up to date, I reinstalled SPICE drivers, virtio drivers, SPICE guest tools and the additional non-guest tools and ...

Performance is unbearable and unusable .. Completely garbage. It's almost seeming like it'll be worth it to go down to 2 monitors temporarily than deal with this.

Before, performance was perfectly acceptable, minimal lag and comparable to any RD connection. After switching to NVidia, input is delayed by at least 5 seconds, if not more.

This goes for Debian 12, which is my typical preferred distro; my 7 cluster Ubuntu k8s group; an outside Ubuntu controller; hell, even Proxmox shell is absolutely unbearable. A CLI !! My god.

Hopefully there is a resolution out there. -- Anyway, just wanted to share for other as they search what I have noticed is a huge problem I have discovered .. Switching from AMD Radeon to NVidia GeForce has caused SPICE to be unbearably unusable. rip.

 

[MrPete]

I know this is old, but just for future reference .. I don't know why or what it has to do with exactly, but I typically have used SPICE viewer without any issue until today.

Today my XFX Radeon RX580 stopped working properly and I was no longer able to display my 4 monitor setup as normal, so I had to switch to a spare NVidia GeForce GTX 1060 6GB. All is well, driver upgrades are completely up to date, I reinstalled SPICE drivers, virtio drivers, SPICE guest tools and the additional non-guest tools and ...

Performance is unbearable and unusable .. Completely garbage. It's almost seeming like it'll be worth it to go down to 2 monitors temporarily than deal with this.

Before, performance was perfectly acceptable, minimal lag and comparable to any RD connection. After switching to NVidia, input is delayed by at least 5 seconds, if not more.

This goes for Debian 12, which is my typical preferred distro; my 7 cluster Ubuntu k8s group; an outside Ubuntu controller; hell, even Proxmox shell is absolutely unbearable. A CLI !! My god.

Hopefully there is a resolution out there. -- Anyway, just wanted to share for other as they search what I have noticed is a huge problem I have discovered .. Switching from AMD Radeon to NVidia GeForce has caused SPICE to be unbearably unusable. rip.

I am curious how this resolved. I suspect a HW issue.