I’ve been messing with homeassistant for a while on the raspberry pi 3, it worked fine, but after a while and many automatons and connections, it began to become slow. I swapped out the sd card for a usb ssd which helped, but that was only a temporary fix. Eventually I got around to buying a cheap intel NUC. It’s a lot quicker. I stuck esxi on it as there’s a free version, and I want to consolidate my growing Pi collection.
Hasio works pretty well on the nuc, its a lot quicker than the pi. However the default disk size is ridiculous. 6gb. It wasn’t long before I ran out of space and had to increase it. This took a ridiculously long time to figure out. I tried to access the underlying OS, as HA runs in a docker container so I cant give it more space from that. After several days of messing around with SSH keys I gave up. I eventaully found a port that said to do this:
- Expand the VMDK
- Boot from GParted ISO (I had to change my bios to “bios” not “uefi” for it to work)
- Expand the partition
- Disconnect the iso, start HA
That was it.
The documentation to install this is somewhat lacking. Assuming you are running Home Assistant as user “homeassistant”, you’ll need to do this:
Connect the camera and enable it as detailed here – https://www.raspberrypi.org/learning/getting-started-with-picamera/worksheet/ don’t reboot it yet.
update the firmware: sudo rpi-update
give the homeassistant user permissions to read the device by adding them to the “video” group – sudo usermod -a -G video homeassistant
You’ll need to restart home assistant to pick up the group changes
Have a look in windows\security\logs\winlogon.log You might have an error in one of your policy settings. Check the file you’re applying permissions to exists and deleting or recreating the rule.
I hate this. I can’t figure out how it works. It will install fine sometimes, and then not others. So far, after various amounts of fiddling I think I have it running. You’ll get the odd error messages, kerberos authentication will randomly work or not… But I think I’m in a happy place now.
Anyway, I plan to add error messages to this post that you may hit. So far, I’ve modified the task sequence so that error 42 is considered a success. This just indicates it didn’t have permission to save the log file for whatever reason. Another one I’ve seen is:
(0xc0002823). Call to function failed. DNS name does not exist. (0xc0000018)
The error code at the end actually means the exact opposite and that there are two IP addresses in DNS for this computer. This happens with us as people often connect via VPN which is in a different range. Delete the offending IP from DNS and try it again.
Had an issue recently where a fresh install of SnapDrive (7.0.3) wouldn’t complete and gave the following error “Product: SnapDrive — Error 1920.Service SnapDrive (SWSvc) failed to start. Verify that you have sufficient privileges to start system services and check your Event Log for more information.”
The server is located in a DMZ on a separate domain and is isolated from the network and the internet. I spent ages looking at this and eventually decided to run a “netstat -ano” to see what the service was doing. For whatever reason it looks like it’s trying to connect to GoDaddy’s CRL server. It seems that as the package is signed by netapp, part of the install downloads a new crl to make sure that it’s valid. Temporarily allowing my host out to the net on port 80 fixed the install issues. Incidentally, as soon as it’s installed you can close that port. From what I can tell it doesn’t check the CRL again….