Behind the scenes of the "move data disk" feature

Installation Home Assistant OS
22

I don’t think you understood how the move data disk feature actually works. With the hybrid mode, your points 1 + 3 + 4 are irrelevant. And for 2, boot times are valid for HA OS only as HA Core is stored on the data disk, not the boot disk.

So please just read and understand what this topic is about. It is definitely not about a migration to a full SSD/NVMe setup.

And fingers crossed for your next OS upgrades - as mentioned, you might want to have a look at the GitHub issues for HA OS (but using an internal NVMe without a third party external case you might be good).

I‘m really curious to hear how things turned out for @rwelsh09 .

23

TL;DR:
I found an xhci_hcd error linked to an incompatible SSD enclosure, I switched to a better SATA to USB and retried. Initially encountering a supervisor error, a reboot resolved it, and the system was successfully up and running 14 minutes after starting the process.

Hello @e-raser & @FloatingBoater

Thanks for your replies. My original data disk was 27GB of my 32GB SD card.

After leaving the system for 5+ hours I finally decided it was time to pull the plug. I connected a monitor to my raspberry pi and rebooted it. The system was unable to start and I eventually had to restore from a backup. But(!) using the monitor showed several errors, the most relevant was an xhci_hcd error which lead me down a path to discovering that my decision to save a few dollars on an SSD enclosure was not worth it. Apparently this does not play nice with RP4. So, I bought this and tried again this morning. With the monitor still connected I initiated the data transfer, this time the HA text said about 10 minutes. I stayed close, watching the monitor, and in less than 10 minutes the system was starting, things were looking promising! There was one small hiccup, I saw something about a supervisor error on setup and couldn’t connect to the system (on port 8123), but I was able to connect to the HA observer (on port 4357, something new I learnt during all this) so I decided a reboot might work and pulled the plug.

Then, in a few minutes, now 14 minutes after starting the move data disk, my system was up and running!

1 Like
24

Thanks for your very helpful experience report :+1:

1 Like
25

“Adding to this thread - I’m recent HA Green user, it worked fine for couple weeks. Yesterday I’ve connected 750GB storage to it and used “Move data disk”, it showed the correct name of the device… but it never booted back up. I connected the disk to my PC and I do see the ext4 partition called hassos-data, so I guess this part work. Disconnecting the disk and power cycling the HA green didn’t start it up properly (it responds to pings though).”

Sounds very familiar :frowning: Don’t know if my report will help anyone or trigger insights, but it won’t hurt, right?

Disclaimer: I’m a Linux noob. I recently got a Raspi 5 with NVMe SSD to mess around with Home Assistant and create a NAS and such. Installed Raspberry Pi OS, switched over to boot from the SSD, got PiMox to work, installed HAOS in a VM, created a 2TB RAID1 NAS with samba from two HDDs connected by a USB dock to the Pi, that I could access from my Windows machine too (piecing together various guides). All in all everything worked quite dandy (on the third attempt).
Then I decided instead of torturing the NVMe drive with the HASS data, might as well use my new NAS and get extra redundancy for free. I couldn’t manage to add the NAS through network, but was able to passthrough the drive directly to the VM. Sure enough, HA recognized it and I naively told it to “move data disk”. It went quite fast since there was not much data at all, but I noticed the folders disappearing from the NAS on the Raspi and it becoming unresponsive in Windows.
I figured it might just need to reboot, so I did that and now the complete system is bricked. :expressionless: Answers to ping, but refuses SSH and VNC. And OF COURSE I don’t have a micro HDMI cable to connect a monitor…

So from my experience, if you’re running HAOS in a VM, don’t move the data disk.
In retrospect I guess it would have been smarter to create the VM on the RAID directly. But as said, half of the time I have no idea what I’m doing.

I’m off on a business trip now while my micro HDMI cables travels to me, in the meantime any ideas for a recovery of the system are highly appreciated. I created a snapshot of the VM before messing around, but even losing the HAOS entirely wouldn’t be a big deal as it’s quite fresh. I just would like to avoid setting up the complete system for the fourth time if I can help it.

Edit:
I think what happened is that I selected one of the RAID drives directly to passthrough to the VM because the proper filesystem was not listed in PiMox. That screwed up the partition table and the Raspi failed to boot because the RAID was mounted without “nofail”.
I booted to an SD card and modified the fstab file on the NVMe to get back into the system. The RAID was beyond saving (at least for my skill level), so I had to re-format it and manually remove the passthrough in the VM configuration because it would keep complaining about not finding the UUID and get locked. Then I was able to rollback to a snapshot I had taken before the debacle and restore one of the HASS backups.
All in all not a lot was lost, but at least I learned a lot.

In a nutshell: Be very careful when playing around with move data disk!