tldr; When TrueNAS boots with both a Windows VM and Plex on startup, network coms get severe interruptions on whatever vlans are being passed to the interfaces on TrueNAS. Doesn't happen when the VM boot up is disabled and is manually started.
The long version:
TrueNAS-12.0-U3.1
I have several VLANs on my home network. The VLANS in question for what's going on here are 5 (services VLAN for servers including file, DNS, DHCP, and AD), 7 (primary client LAN), 20 (IoT), and 40 (Media devices including streaming devices and Plex jail).
In TrueNAS, I have Plex installed (current version) and a VM running Server 2016 as a 2ndary DC/DNS/DHCP. I have 2 different NICs i've tried this on. 1 is a 4 port i340-t4 and the integrated RealTek gigabit nic.
When I used the RealTek NIC, I configured my switch to pass 5, 7, 20, and 40 VLANS with 5 as the native for the port. Whether I do manual or DHCP, I'm getting and IP in that VLAN in TrueNAS and that part is working fine. I then went on to create the VLAN interfaces in TrueNAS for 7, 20, and 40.
I assigned the Plex jail to use vlan40 interface and DHCP. Start up Plex, everything's fine, I'm getting the right IP assigned to it and all looks well.
I assigned the VM to use RealTek NIC and tried both VirtIO and Intel adapter types, as soon as the VM starts, Plex looses network connectivity and my devices in VLAN 40 start getting IP assignments from VLAN 5. As well, all devices in any of the VLANS being passed to TrueNAS start dropping packets.
So I changed things up, threw the 4 port NIC in there. moved TrueNAS's IP to igb0. Assigned Plex use of igb2 and the VM igb3. I deleted the vlanxx interfaces I created and changed my switch config to only pass vlan 5 on 2 ports and vlan 40 on one port (using 3 connections at this point).
So fired up Plex, getting the right IP again. Fired up the VM, it looks fine. Let everything run for a couple of hours and made sure both DCs are talking and replicating with each other. All seems great.... until... dun dun dunnnnn
I rebooted TrueNAS just to make sure everything was going to come up properly and neither Plex nor the VM had network connectivity and everything in VLANs 5 and 40 start dropping packets again and my streaming devices start getting IPs in vlan5's subnet (should be 40's).
So after some process of elimination, I landed on disabling the autostart of the VM and leaving Plex's autostart on. When I reboot like that, Plex starts up fine, everything on my network is behaving, and when I fire up the VM manually, everything is good.
I'm at a loss here. Why would starting the VM after TrueNAS is fully up make a difference in this scenario? Is there any way I can add like a 60 second delay to starting the VM if the root cause can't be fixed?
The long version:
TrueNAS-12.0-U3.1
I have several VLANs on my home network. The VLANS in question for what's going on here are 5 (services VLAN for servers including file, DNS, DHCP, and AD), 7 (primary client LAN), 20 (IoT), and 40 (Media devices including streaming devices and Plex jail).
In TrueNAS, I have Plex installed (current version) and a VM running Server 2016 as a 2ndary DC/DNS/DHCP. I have 2 different NICs i've tried this on. 1 is a 4 port i340-t4 and the integrated RealTek gigabit nic.
When I used the RealTek NIC, I configured my switch to pass 5, 7, 20, and 40 VLANS with 5 as the native for the port. Whether I do manual or DHCP, I'm getting and IP in that VLAN in TrueNAS and that part is working fine. I then went on to create the VLAN interfaces in TrueNAS for 7, 20, and 40.
I assigned the Plex jail to use vlan40 interface and DHCP. Start up Plex, everything's fine, I'm getting the right IP assigned to it and all looks well.
I assigned the VM to use RealTek NIC and tried both VirtIO and Intel adapter types, as soon as the VM starts, Plex looses network connectivity and my devices in VLAN 40 start getting IP assignments from VLAN 5. As well, all devices in any of the VLANS being passed to TrueNAS start dropping packets.
So I changed things up, threw the 4 port NIC in there. moved TrueNAS's IP to igb0. Assigned Plex use of igb2 and the VM igb3. I deleted the vlanxx interfaces I created and changed my switch config to only pass vlan 5 on 2 ports and vlan 40 on one port (using 3 connections at this point).
So fired up Plex, getting the right IP again. Fired up the VM, it looks fine. Let everything run for a couple of hours and made sure both DCs are talking and replicating with each other. All seems great.... until... dun dun dunnnnn
I rebooted TrueNAS just to make sure everything was going to come up properly and neither Plex nor the VM had network connectivity and everything in VLANs 5 and 40 start dropping packets again and my streaming devices start getting IPs in vlan5's subnet (should be 40's).
So after some process of elimination, I landed on disabling the autostart of the VM and leaving Plex's autostart on. When I reboot like that, Plex starts up fine, everything on my network is behaving, and when I fire up the VM manually, everything is good.
I'm at a loss here. Why would starting the VM after TrueNAS is fully up make a difference in this scenario? Is there any way I can add like a 60 second delay to starting the VM if the root cause can't be fixed?
