Dear God, please help!


Install in an Ubuntu image under windows , you can then have a raid , system shots , backups every once in a while , to external disk , flash or whatever , if it dies just restore backup , whatever windows machine will work , wife will be happy !!

I had some serious issues with RPi 3b+ with a lot of hangups. From what I could read from Glances it was related to I/Owait.

Found this definiton of it:

For a given CPU, the I/O wait time is the time during which that CPU was idle (i.e. didn’t execute any tasks) and there was at least one outstanding disk I/O operation requested by a task scheduled on that CPU (at the time it generated that I/O request).

Moved to an RPi 4 and all my issues was gone instantly

UPDATE 05/07/2020

So, I have tried everything to get Ubuntu Server on this Lenovo ThinkCentre M72e with no luck. UEFI enabled or disabled, Drive mode set to IDE or AHCI, Fast Boot/QuickBoot enabled/disabled, it all makes no difference. Additionally, switching to Ubuntu Desktop gives the same results. While the Live CD will boot and the install will complete successfully, “No operating system found” is returned upon first boot after installation. GParted shows the boot volume is flagged properly, yet the system apparently is biased against everything but Windows. I really didn’t want to have to lose the drive space and resources Windows will take just to run this as a VM, but that’s looking like what I’m going to need to do.

The problem does not appear to be limited to Ubuntu. Someone tried to instal libreELEC on the M72e and encountered exactly the same problem you did.

Ultimately they had to alter the BIOS in order to correct the problem. It’s a long thread describing the many things that failed to work and culminates in a post where they altered the BIOS. A subsequent post explained that the problem was in the latest BIOS (at the time) and the solution was to simply revert to an older version.

Ok, through trial and error and a BUNCH on help here and a number of DMs from @kanga_who, I’m finally up and running on the ThinkCentre M72e with Ubuntu Desktop 20.04 LTS 64-bit.

Here are the BIOS settings I had to get in just the right combination in order to get this working:

CMS - No entry in BIOS
Boot Mode: Legacy
Boot Priority: Legacy
Boot Order: HDD first, everything else disabled.
Quick Boot: Disabled
Drive Mode: IDE
Drive Native Mode: Disabled

Compared to my old RPi 3B+, this [email protected]/16GB/500GB tiny system SCREAMS! This was a refurbished unit I picked up from for about $120USD. Thinking back it’s actually less than I’ve put into my RPi 3B+ running a 64GB MicroSD card (and other crashed cards), case, case fan, and upgraded power supply. All headaches aside, it was totally worth it to get this thing going.

All I need to do now is get my Z-wave devices paired with my new Z stick and rebuild my all of my old yaml automations in NodeRed and I’ll be all set to pull the plug on the old RPi 3B+! I can’t wait!

Thanks all!

1 Like

I don’t get it. Those are all fairly vanilla BIOS settings. What were they set to when it failed to boot?

Did you have to re-install ubuntu after making those changes?

I did, but after the reinstall it came up properly.

Thank you so much for providing those guides! I had difficulty with 20.04 and spent several hours banging my head against the wall yesterday. Started fresh with your guide and mostly working within an hour. I do have one roadblock and not sure if you or someone else here can help. My particular install is using VirtualBox on windows. I have the network adapter in bridged mode and a private reservation on my router. I can SSH, get to files via SAMBA, verify docker containers are running in Portainer web ui, and get to HomeAssistant.local and completed several initial integrations/devices and can even connect with the mobile app on android.

The issue I have is that in the HomeAssistant.Local web interface clicking on the Supervisor menu item results in the web interface freezing. A chrome trace shows that calls to these three URLs get stuck in pending and never return a result.
http: // homeassistant.local:8123/api/hassio/supervisor/info
http: // homeassistant.local:8123/api/hassio/host/info
http: // homeassistant.local:8123/api/hassio/core/info

I’m not exactly sure where to look for logs and the logs I have seen aren’t showing errors. Docker is showing Supervisor as running and SSH ha cli shows supervisor as on.

I am new to docker, virtualbox, ubuntu, and home assistant and appreciate any pointers. Since Wink is attempting their extortion scheme it’s time to move to something better just need a little assistance bridging the gap.

The suggestions in this thread need revision.

“Revision”. :slight_smile:

I’ve played with this further and found that I can get Supervisor to load from a web browser on the Host OS. Just not from another machine on my network. Tried running HassOS VDI in VirtualBox and same behavior whether VirtualBox is network is configured as NAT or Bridged. Not sure why Home Assistant and Lovelace will work on non-host os but Supervisor won’t.

OP back again! My Lenovo is still running strong and super fast, I love it! Then I saw the post about sun-setting the supervisor installations and was bummed. Then read that the sun-setting was put on hold and was happy again!

Well, as it turns out, I was gifted/stuck with a Series 7 DP710A4M that had a bad display. A bunch of colored lines running from top to bottom but the screen was still usable, just ugly. Well $68USD later and a trip to Ebay, the replacement screen arrived today and I got the bad one swapped out and viola!

New HA Touchscreen system!

It’s a 7th Gen i5, 12GB, 1TB, but it’s upgradable!
So, I upped it to 16GB, stuck a 250GB m.2 drive in it, disabled SecureBoot and Quick boot, and set the UEFI to UEFI Operating system and installed Ubuntu 20.04 desktop and everything works out of the box! I’m so stoked! Wifi - Connected and working, Touchscreen - accurate and working, Ubuntu booted after the first install attempt - WORKING!!!

Now I just need to follow the guides posted above to get HA installed and move over my Z-Stick!

Thank you to everyone to assisted with my UEFI / HA nightmare, especially @kanga_who. Such a great community!



I got my m72e working finally by upgading the bios - for which I had to take the lid off and move the clear-cmos jumper and boot from a DOS usb stick (freeDOS installable from Rufus) and fun the DOS based BIOS updater.

Now playing with Proxmox.

1 Like

Just finished moving my daily driver install to Proxmox today. The learning curve was a little steep, but now that I understand it, I am coming to like it. Currently running 3 VMs.

  1. Home Assistant OS (Google drive backup, VS code, AdGuard, Chrony, zigbee2mqtt, Unifi controller, Tasmo admin, SHH - HA add-ons)
  2. Ubuntu server 20.04 (MQTT server, Plex server, Portainer, HA-dockermon, OVPN, Tranmission - all using Docker)
  3. Ubuntu 18.04 (Shinobi for CCTV)

Don’t forget to enable regular snapshots. Save them on another host/drive/…
Once or twice when the home assistant supervised backup failed on me I found it very useful to have backup of the VMs!
You can also for some other services (eg. pihole) run them of a LXC container.
I have been running proxmox 4 then 5 on a nuc i3 and it is really good.

I am stubornly refusing to do this… for now… I live in hope the madness of depreciating the generic linux install will be revoked. This despite having a chinese XCY mini PC running Proxmox with a bunch of VMs that I fire up only when I want to test something. I have multiple variants of HA installed on it but I REALLY REALLY don’t want to have to reconfigure my Lovelace and sensors etc to pull data from a different VM and deal with the BS of different network component names etc on the different hardware - it’s just a mega PITA and I’m not going to sink a day into doing that when there is the least chance I won’t have to.

I thought the same, but, with my business still closed because of COVID and time on my hands I figured it was a good time to make the change and learn so I don’t have to potentially do it down the track.

I found that as soon as I passed-through and mounted my CC2531 to the HA VM, it was seen on next boot and zigbee2mqtt worked as before with no need to re-add devices to the zigbee network.

As far as sensors/switches go, I gave the VM running my mqtt server the same IP as my old supervised install, so everything worked without issues.

I’m actually finding that Shinobi is working better in its own VM and reboot times for HA are faster. Go figure.

I have just ordered a 16gb RAM kit to replace the 8gb I have in the Dell Optiplex 990 I have running everything. This is so I can give each VM 4gb instead of 2gb. Other than that, all is working very well.

edit: The other thing worth noting is that you can give a VM any MAC you like, so if you don’t have to change anything in your router, just give the VM you want the same MAC as the machine you used to use.

1 Like

The most “generic Linux” install method is the Python virtualenv, and I don’t think anyone is suggesting that method is going away. HA is just another thing that runs on my Pi, and I love having it that way.

The Linux generic install has a very specific meaning in the docs and has nothing to do with your definition. In any case, HA Supervised replaces that install method now.