I assume you are running your own DNS here? Wait no those are off subnet. The reason you’re overriding your router for DNS or us that just what the router is passing on. Each mfr does this differently.
Is pingable until “Host unreachable” after shutdown. So no double ip adres.
1 Like
It’s just passing through the dns servers of the router.
1 Like
Just wondering… The fastest way to see this… You wouldn’t be able to operate Wireshark would you - man I’d really love to see whats going on on the wire when that VM starts.
This thing running Unraid is wired directly to one of the router ports isn’t it?
Pinging FROM?
(including type of host, and IP settings)
I just installed wireshark on unraid (the App Store is fab). But I don’t really understand what’s going on 
. Unraid is via a switch connected to the router and I’m pinging from the unraid machine to the virtual machine.
1 Like
Running only HA on a VM or Container also makes no sense. Life is so simple with HAOS on bare metal in my server closet. IP and COM ports just work.
Wireshark is wonderful but I dont have time to talk you through the config and you can absolutely expose bad stuff on your network by randomly posting captures it’s engineer level tooling (read: absolutely great tool to learn - not for this)
OK then you’re pinging from here:
I need to know it’s IP address and subnet mask.
That confirms you’re pinging from unknown IPV4 - V6 is disabled on the source.
Then you said:
I think we’re all making an assumption here. Are they marked unavailable in HA or are they unable to be pinged from anywhere else on the network? If they’re tuya wifi devices you should be able to ping a few IPs according to whatever is giving out DHCP from the Unraid box and from a client NOT on the Unraid VM system. Have you tried? (This one’s important as to where the issue is)
Edit: additional question
and
Says you are dynamically assigning the HA IP address via DHCP - at the router? - Are you somehow reserving the HA IP address in your router? (DHCP server)
Like they do on a hypervisor. You’ve said multiple times that you have no interest in virtualization or containers - that’s fair enough - but you really have no clue what you are talking about and you shouldn’t give advice on something you have proclaimed you have no interest or knowledge on. OP never said anything about ONLY running HA as a VM, that’s something you assumed.
The way you ask twice why the OP has HAOS as a VM, and respond "Features are not a “why”. when you get presented two (of many more) great reasons of running it virtualized is just ignorant. 
The have been a post or two about issues with the deco router on the forum with a similar issue.
Try to search for the router name.
1 Like
I knew I saw it somewhere Wally… I think ops reservation is failing to re-assign IP addresses at renew and to Op. It looks like the Tuya devices drop where in actuality it’s HA failing to re IP.(because Deco?)
I would:
- assign a fixed IP to HA (outside of the DHCP range)
- figure out the IP addresses from my devices
- with ping establish which devices specifically are falling of my LAN (and see if there is a difference in behavior between wireless and wired devices)
- And if possible: check the logs from my AP
1 Like
I can’t find the thread at the moment, but the description of the error was pretty much identical and it was the solution was something with the deco router. Maybe update or roll back the firmware?
Was that the one with the weird MTU setting?
No answer but for what it’s worth, I run HAOS on an Rpi and another copy in vm on a laptop. IP addresses for both are hard allocated on the tplink deco. The Rpi runs without fault but the vm does have occasional issues with loosing network connection. It used to happen regularly but has been good for the past few months. No idea why. I did find a discussion on this linked below but no resolution. I treat the vm copy as my test environment and it’s faster than the Rpi.
For general network troubleshooting I find wifiman on an android phone very helpful.
Unraid ip is:192.168.1.221 and 255.255.255.0 the submask
The devices are marked unavailable in HA and are not pingable from any device. To be clear: the vm with HA is perfectible reachable, just as any other wired device.
I’m reserving the IP-adres at the router.
Good suggestions, I’ll try them first time the network is failing!
It happened again… I did assign a fixed IP (outside the DHCP range). Updated to HAOS 13.1. I looks like all my 2.4Ghz Wifi devices fail. My 5Ghz devices do work…
Some dual band routers try to push devices from 2.4Ghz to 5Ghz by kicking them off the AP’s 2.4Ghz band and refusing them back on that for a short period to force them to try 5Ghz, but not all devices can connect to 5Ghz and do not get the hint and then just give up trying to reconnect.
1 Like
But when I reboot the vm everything works again 