"
Hi all,
I've been reading a lot about the issues between Hyper-V and BSD with regards to networking performance. RSC, not RSC, etc.
Right now I have tested the following and have found no solution:
Host1:
AMD CPU
Windows 10 22H2 up to date with patches
Intel NIC with updated drivers
Host2:
Intel CPU
Windows 10 22H2 up to date with patches
Intel NIC with updated drivers
In both hosts with either G2 or G1 hw versions, a fresh install of OPNSense with only 1 network adapter, manually configuring an IP and setting the gateways.
Internet -> OPNSense -> Host OR other device == ISP max speed
Host or other device -> OPNSense -> Internet == 0.1Mbps or less
I have seen several posts about previous versions of OPNSense and the problems between rsc at the adapter and vSwitch level. I have disabled it, disabled VMQ and other performance enhancements, ENABLED offloading. Disabled level 2 offloading at the driver level.
Nothing has made any difference.
Since it has been so easy to replicate I wanted to post it here to see if anything else can be diagnosed or collected to help fix it someday.
I do not want to change the hypervisor because the machine where I want the OPNSense VM has other roles for my home network and I can't pass through the integrated GPU to a Windows VM if I wanted.
I've been reading a lot about the issues between Hyper-V and BSD with regards to networking performance. RSC, not RSC, etc.
Right now I have tested the following and have found no solution:
Host1:
AMD CPU
Windows 10 22H2 up to date with patches
Intel NIC with updated drivers
Host2:
Intel CPU
Windows 10 22H2 up to date with patches
Intel NIC with updated drivers
In both hosts with either G2 or G1 hw versions, a fresh install of OPNSense with only 1 network adapter, manually configuring an IP and setting the gateways.
Internet -> OPNSense -> Host OR other device == ISP max speed
Host or other device -> OPNSense -> Internet == 0.1Mbps or less
I have seen several posts about previous versions of OPNSense and the problems between rsc at the adapter and vSwitch level. I have disabled it, disabled VMQ and other performance enhancements, ENABLED offloading. Disabled level 2 offloading at the driver level.
Nothing has made any difference.
Since it has been so easy to replicate I wanted to post it here to see if anything else can be diagnosed or collected to help fix it someday.
I do not want to change the hypervisor because the machine where I want the OPNSense VM has other roles for my home network and I can't pass through the integrated GPU to a Windows VM if I wanted.
