HASS.IO -> transfer from SD card to SSD or USB

I need assistance if anyone can help. I migrated to SSD due to failed SD card and did a new install. I am on HassOS 5.2 and 0.116 and the new installation is working perfect. I had google backup add-on and have 4 snapshots, and been trying to restore, but keep getting the following error.

2020-10-08 11:03:50 ERROR (MainThread) [homeassistant.components.hassio.http] Client timeout error on API request snapshots/new/upload

I recall that I had set the snapshots to be password protected, so maybe that is the problem?

What version were you on when you took those snapshots? Have you considered rolling back to that version and giving it a shot?

got that same enclosure today (that’s what I would say based on the information that I saw on amazon). Just installed it. Will report on what happens now. I so much wish this solves my issue…

@gremblin Could you share some pictures? Are you monitoring the temperature on the SSD somehow?

The last snapshot was on ver 0.115.2. I have no clue on how to roll back to 0.115.2. Any pointers? Would try anything to get back as I have 6 months of work on my HA.

Try this: How to downgrade Home Assistant?

Thank you @kitus. I managed to change it to 0.115.2. However, on the Snapshot page, at the 3 dots on the right corner, only shows the reload option, the upload option is not shown (it was there in 0.116), so there is no way I can add the older (my last working snapshot) to the list of available snapshots. Even the Google Snapshot does not show the older snapshot and pressing “Sync Now” in settings does not update the list, even though the snapshots are on the google drive. I just can not seem to be able to add the snapshot to the list so I can try and restore it. Very frustrating.

@Sunil have you tried to use the Samba Addon to upload your snapshot? It should be uploaded to the backup folder…

@Sunil, Erik is right. Try that and let us know!

@erik73 and @kitus Thank you. Installed Samba, opened the config directory on my computer, there was no backup folder, so I created one and copied the snapshot in it. Restarted HA and still can not see it in the Snapshot page in HA. Any other suggestions?

You should see these folders

Thank you very much @pepe59, @kitus and @erik73. With you help I managed to copy the snapshot to the backup folder and restored the backup. Restarted and a few things broke, but managed to fix most of it and updated it to 0.116. Greatly appreciate your kind assistance. :grinning:

I am not monitoring the SSD temp with any sensor, I just felt it with my finger and it was “hot.” I also had my rPi and SSD in an old alarm enclosure without any venting or fans. When closed and locked, I noticed that the rPi CPU temp was much higher with the SSD. This I am monitoring with the system monitor component. The temp increased about 50°F, so I know that SSD was kicking out a lot of heat. I purchased these small heatsinks with some double sided thermal conductive tape and installed them on the drive. The SSD still felt hot when I tested with my finger, so I added an external fan that I had laying around… its a 12v 120mm fan I had from an old computer build. I have a 12v power rail in my enclosure, so I just tapped into that and zip tied it to the case… It’s super ghetto, but it was just to test to see if I added some cooling if it would run stable. I plan to redo the case design to make it more permanent, now that I feel like this was the issue. I have also purchased a better SSD to see if it will run cooler. I am currently running on this USB jump drive. After having the suspicion that it was running too hot, I ripped off the nice aluminum enclosure and added the heatsinks to the chips. SanDisk claims that the drive is using SSD chips, but I have a feeling that it runs hotter than a NVMe SSD and possibly hotter than a standard SATA SSD. I can post a photo later if you wish, I can’t access my system right now due to a sleeping baby.

@gremblin I need to review in detail your post. Thanks a million. Today my system went down for 1h and I very tactically enabled a few additional system monitor sensors (I was not monitoring CPU temp).


It turns out that CPU temperature definitely skyrockets right when the system was booting up. That’s expected due to the load that the booting process entails.

This is the load of the CPU


Unfortunatelly, I don’t seem to have insights prior beyond when the system came back up unless I use some sort of influxdb and grafana.

Should I give a try moving back to a SD card to tell apart whether all my problems are SSD related? what do you guys think?

does the HDD Tools add-on create this entity? I’m also trying to run that addon likewise @engpedroamorim but the output that I see on the front-end is full with:

[Mon Sep 28 16:12:53 WEST 2020][Info] Sensor value: °
[Mon Sep 28 16:12:53 WEST 2020][Info] Sensor update response code: 400

My raspberry went down 3 times already today. I was hoping that a new enclosure would remediate the problem but it clearly is not the case. Any call to action that you guys could think of, would be extremely appreciated. I’m more and more considering reverting back temporarily to a mundane SD and see if the system works more stable.

The CPU load that I shared above worries me. Also my RPI has all the USB ports in use with RFXTRX, CC2531, Z-WAVE and the SSD. This could also be making the RPI choke, right?



Yup. Basically, what you are seeing is the rPi failing to have enough power for all your USB attached devices and the rPi itself. A simple solution is to get an externally powered USB hub and attach your devices to that and then connecting the hub to your rPi.

You should be able to witness this behavior by running vcgencmd get-throttled and/or vcgencmd measure_volts on your rPi.

I currently have a rPi3B+ at a friend’s house with 12 external SDD’s attached to two externally powered USB hubs as a makeshift NAS. It has never throttled once. :slight_smile:

@code-in-progress I’m running the former hassio. Can I still run those commands? Also, I must say that my system works well for several days in a row, then it collapses.

If it did not have enough energy to power all the peripherals attached to the USB ports, it should not even be able to boot, or it should fail more often, right? In other words: what is it in the snapshots above that makes you think that the system is short on power?

thanks a million!

Honestly? I’m not sure. You could try them with the SSH addon and see.

Not necessarily. Basically, the rPi will try to provide as much power as it can to peripherals and by doing so, it will throttle itself to achieve that. The power draw largely depends on how much each device is accessed and not all USB devices draw power the same way. Sometimes, the Pi will fail to boot if there is too much power draw on the USB bus (or combination of the USB bus/network bus on the rPi3B+). But that’s not always the case.

Found a way to run the two commands

# docker container exec homeassistant /opt/vc/bin/vcgencmd measure_volts
# docker container exec homeassistant /opt/vc/bin/vcgencmd get_throttled

what do you think?