At this point, I'm just trying to determine some general thoughts and direction.

 

I am consistently getting reasonable ping and up/down speeds on my unRAID box itself.  Ping for the unRAID box itself is ranging from 15ms to 90ms typically, which is reasonable.  Other machines on the LAN are getting similar results.

 

I have set up a Windows 10 VM so my son can play windows games on it using GPU passthrough.  In general, this has been very successful.  However I am seeing very erratic ping behavior, even when the unRAID box is not otherwise under load.

 

Ping for the Windows VM have surged up and down into the 300ms range and are much more erratic.

 

Is this a typical result for VMs in general?  Is it related to how the shared NIC prioritizes traffic?  Are there settings I can change to alleviate this issue, or should I install a second, dedicated NIC to pass through for the VM?

 

Any advice is appreciated.

 

 

Edited October 21, 20178 yr by HoustonDave

8 minutes ago, HoustonDave said:

Ping for the unRAID box itself is ranging from 15ms to 90ms typically, which is reasonable.  Other machines on the LAN are getting similar results.

 

Are you using wireless? Wired ping should always be <1ms

1 minute ago, johnnie.black said:

 

Are you using wireless? Wired ping should always be <1ms

 

Apologies, I should have been clearer.  I am pinging from the device/VM to an external IP (example:  8.8.8.8) not pinging from one device on the LAN to another device on the LAN. 

 

Yep it is wired connections.

Yep, I suspected that after posting, a little more latency is normal, but I never experienced nothing noticeable, just tested 8.8.8.8 and got 37ms (VM with passtrough NIC) vs 56ms (VM with virtual NIC)

Thank you.   That's a helpful data point on whether a simple change from virtual vs passthrough NIC makes a significant difference in general.

Just as a quick update, it appears that the issue was not the VM, and was the WAN instead.  I was just having good/bad timing on when I pinged from various computers so it appeared the issue was limited to the VM.  Some reconfiguration of my WAN connection to modem seems to have resolved it.