ucliker

I think adding the correct metrics worked. Since my 10gbe card is directly connected to another card in my workstation unRAID sees that and assumes it needs its own default route. I just set that to metric 2 and my default gateway (192.168.1.1) to metric 1 and it seemed to fix everything. Thanks @bonienl for comment regarding the router setup, that helped. Also thanks @ken-ji for the help with the metrics, the policy-based routing definitely made more sense once I read about it on the Cisco website. I may have had metrics setup prior to upgrading at some point and never changed it back after swapping nic cards or something.

@bonienl I am not configuring anything. Every time I reboot the systems sets a default gateway like its supposed to. It only sets one but it is the wrong one every time. I then need to delete that one and add the correct one. It using my 10gbe adapter (eth2) which isn't connected to the internet instead of using eth0.

@ken-ji I never had to change the metrics in the past, I always left everything as the default (1). I'm just surprised that after all these years of using unRAID that this is happening. Its not creating two default gateways when I reboot its assigning the gateway from eth2 as the default. I guess I will set the metrics and see what happens.

From my experience the bridge label always matches the interface label, eth0 = br0, eth1 = br1, eth2 = br2. I have had so many networking issues like this since the last release of unRAID and I thought upgrading would fix it. For me the main issue is my default gateway changes to the wrong gateway every time I reboot now. My 10 gbe card isn't connected to the internet but it always uses that as the default gateway now. I know I can write a script to add the correct gateway but I would just like to know why.

Check your clock "Clock Unsynchronized" This may not be your issue but I had it in the past and it is listed as an error in your logs.

I figured it out finally. I think it was my router since I had my static IP Address set to my old address but once I changed it everything was fine I think. So earlier today I deleted network.conf and rules and changed the eth0 address to a new address but the old address still works?

tower-diagnostics-20190516-1855.zipThanks, I was trying to figure out how to get the diagnostics file. Here is my diagnosticstower-diagnostics-20190516-1855.zip files.

For the past few days I have been having network issues with my unraid server (just the internet access to the community apps). Today I decided to update to 6.7.0 from whatever version was prior to this one. Now I cannot access the GUI at all from any device or browser. I tried deleting the network.conf and rules since I wanted to change the IP Address anyway but I still can't access it. It works fine with SSH and direct connected to a monitor. Any help would be greatly appreciated.

Is your Rider config inside the default config of letsencrypt? or are you using something else? I don't use Rider but when I created my Seafile config for letsencrypt and nginx I had to create a separate config file just for Seafile and setup my proxy as seafile.DomainName.com. You should be able to reverse proxy any docker container although it may take more tweaking for Rider. I have Pycharm and if I get that to reverse proxy it should be similar, I will let you know.

I think you’re correct. I was testing blender at the time in that VM and then Ghost Recon wildlands and it had huge performance loss. I tested with a GTX 1060 and Quadro k4200 and those saw no performance hit at x8.

I had terrible performance with x8 but only when I was running my Titan X and not my old Quadro. I do agree Q35 machine type is better even for MacOS vm’s.

Can you check and make sure the 1070 is running at x16 and not x8?

My guess would be driver issue. Did you check the power supply and cables?

When I was running a similar VM and switched hardware I had to create a new VM from scratch. You may want to also uninstall the Nvidia drivers in the VM and then reinstall them. Also check the pcie slot like 1812 mentioned but also check the power cables etc. another thing would be to check the power mode for the GPU in the nvidia setting menu.

I have had one die on me but it was always running extremely hot. These cards are used enterprise cards that got hammered for years so them dying is expected.