Last few updates I haven’t been able to update hassio core past 114.0…from the GUI it briefly loops then goes back to available, from command line it prints success but doesn’t actually DO anything.
I’ve tried reboots of HassOS and it doesn’t seem to make a difference.
The system log just prints “ERROR (MainThread) [supervisor.utils] Can’t execute update while a task is in progress” but I have no idea what “task” it would be doing?
Anyone have ideas?
I get that error whenever I get impatient with an update and try clicking update a second time before the first update completed. Weird it’s doing that after a reboot though. Hopefully you don’t have a zombied update flow or something. First, take a snapshot (partial, no database addons). Next, I would first try waiting an hour or 2, and try updating again. If it still fails, you may have to R&R HA to get it fixed, then use your snapshot to restore things back to normal.
I’ll try making a snapshot but I don’t know what that would do.
What’s “R&R”? I do hope that’s not wipe and reinstall everything…that would REALLY suck having to start over.
The snapshot will get you going exactly as before after a reinstall.
since reinstall for hassos would wipe the disk, how would I get it back though?
Make a snapshot, download snapshot, wipe & reinstall, go through on-boarding (don’t need to bother with any configuration, just make an account so you can login), install SAMBA add-on, upload snapshot you made previously, restore snapshot. At that point you should be back to exactly where you were when you made the snapshot.
Hmm…backed up my VM and attempted to restore snapshot to a new one (wow, they have KVM/QCOW images now!) after a lot of pain getting the network back up on the VM (grr, we really need a network config UI for HassIO, I hate nmcli) I think everything except MariaDB is working.
Is there any way to get my history back, or “recover” from the supposedly corrupt restored DB?
Unrelated, very annoying Firefox keeps trying to change HTTP to HTTPS and making me think the homeassistant machine is broken when its reloading…UGH
Hmm, you mentioned MariaDB, is that the historical DB that you’re using instead of the built-in HA one? I don’t use it, but I think I remember reading that it doesnt do live backups well, and you might have to shutdown that addon before starting the snapshot for it to restore right; but that might be outdated information, I’m not sure :/
Yes, it seemed like the performance was really bad with the default sqlite and as much as I don’t like the complexity going to mariadb seemed to make the whole system much more responsive especially viewing graphs.
I’ll try shutting it down and re-backup and re-restore…whole thing makes me wonder how much I really wanted to update…
Looks like stopping Maria before snapshot and stopping Maria before restore worked. I think its back alive.
I also notice the disk space is ~14GB used instead of ~47GB used…no idea why.
Guess I’ll find out in the morning if my automation work…
Good you found that you could stop db’s before backup if you want to preserve that database… forgot to mention this option. What version did you update from? IIRC, one of the recent versions came with some major changes to how databases are done; the update would ‘refactor’ the db or something like that, making it much smaller than before.
From 114.0 to 114.4
It hadn’t been letting me do any upgrades (or even reload without host power cycle) since 114.0 always saying success but then not doing anything.
I don’t think its fixed though.
I still can’t restart homeassistant:
20-08-29 04:42:51 INFO (SyncWorker_4) [supervisor.docker.interface] Start homeassistant
20-08-29 04:42:52 INFO (MainThread) [supervisor.homeassistant] Wait until Home Assistant is ready
20-08-29 18:58:13 ERROR (MainThread) [supervisor.utils] Can't execute restart while a task is in progress
Hmm, I’m pretty sure it was .113 that had the db change out. The process could take some time, and there were lots of warnings in the release notes about giving the db time to redo itself… a decent number of posts from users who had issues with the db refactor as well. I wonder if you may have also had a botched db upgrade from before?
If so, you might try it again without the old database (partial snapshot w/o any db’s). Many folks say they don’t care about losing info on when doors were opened/closed months ago, but everyone’s use case differs. I know, it can also suck to lose several years worth of influxdb depending on what it’s used for, but if the data can’t be saved, it is what it is. There is certainly something to be said about offloading the db to a separate server (nas or similar). Then you don’t have to worry about an HA restore glitch hosing any of that data (but then there’s 2 servers to maintain as well).
Yeah, mostly I don’t want to lose stats “at this time” because one of the things I track is internet ping, DNS, and speedtest and I’ve been having issues recently trying to keep tabs on my ISP the past couple weeks. I wouldn’t care as much otherwise.
I’ll try removing the DB and see if it gets un-stuck but I’m really confused what it’s hung up on…and there seems no way to tell!
I just tried removing all my addons except maria, deconz, and SSH…stopped those addons and disabled on boot, cleared out all but the HTTP port/interface for the configuration.yaml and still seems stuck.
I guess maybe I will have to re-reinstall and try to restore one thing at a time to try and tell when it breaks, no idea what else to do.
FWIW, I have an asus router that does all that stuff (and then some) using scripts with merlinwrt firmware. There are other routers out there that can run user scripts like that, not just asus. I figured I’d mention this in case you have a merlin compatible router. Then you could offload those tasks to the router, and easily add more security/privacy/features to your network while you’re at it.
using pfSense on a x64 mini-PC with server NIC as a router, but I can’t get all the stats I want. I want it to be able to log every ~10 seconds and see if my core switch, router, modem, external connectivity, and DNS at different levels, plus running speedtest.net test hourly (been tracking speed issues, ISP always blames “my router” or “my modem”…I’ve seen it blip and my modem is still reachable as is my router). If I could figure out how to log all the modem signal levels I’d be doing that too.
Hassio gives better flexibility.
Ok - I think I’ve tracked the issue to one line in my config file…but I don’t understand why. Maybe its a bug?
http:
# server_host: 192.168.1.221
server_port: 80
Commenting out that line fixed my issue…but now it’s accessible from an untrusted interface of IoT stuff. I don’t want the login page accessible on any other interface.
Is there some firewall I’m unaware of?
I’ve got HassIO sitting on 2 VLANs with their own NIC on each:
192.168.1.0/24 – my tursted machines (phones, laptops, etc), Internet access, VPN allowed to log into systems on this segment, etc)
192.168.3.0/24 – untrusted IoT stuff, no Internet, no access to any other systems (but HomeAssistant needs to sit on here to have mDNS work for ESPHome and other device discovery to work right – when I tried to have it go thru the router to get to this VLAN it couldn’t find any of my devices/integrations)
The server_host: 192.168.1.221 has been working great since I first set this up for making it only seem to bind the UI on the one interface only, but then it inexplicably breaks the ability to update or restart HomeAssistant without rebooting the whole host?