trying to upgrade to 92.2 from hass.io panel.
getting this in my log.
19-05-03 08:13:36 INFO (MainThread) [hassio.homeassistant] Update Home Assistant to version 0.92.2 19-05-03 08:13:36 INFO (SyncWorker_1) [hassio.docker.interface] Update image homeassistant/tinker-homeassistant:0.92.1 to homeassistant/tinker-homeassistant:0.92.2 19-05-03 08:13:36 INFO (SyncWorker_1) [hassio.docker.interface] Pull image homeassistant/tinker-homeassistant tag 0.92.2. 19-05-03 08:13:38 ERROR (SyncWorker_1) [hassio.docker.interface] Can’t install homeassistant/tinker-homeassistant:0.92.2 -> 404 Client Error: Not Found (“no such image: homeassistant/tinker-homeassistant:0.92.2: No such image: homeassistant/tinker-homeassistant:0.92.2”). 19-05-03 08:13:38 WARNING (MainThread) [hassio.homeassistant] Update Home Assistant image fails
Anyone seeing issues with HA 0.92.2 becoming unreachable and going offline? If I reboot the host, it runs for about 12hrs and then I notice that Xiaomi motion sensors stop turning on lights. Go to the UI, not reachable. Samba and SSH OK and when I run a check through SSH it reports back OK.
I was having issues with MariaDB which I thought was the issue (Synology and local add-on) but now I’ve returned to the local dB for a while and it’s doing it with that now too. Yesterday it did a very weired lockup where all of the Internet based elements (Darksky, Tides, Nabu Casa Cloud etc.) were unreachable. Again, a host reboot sorted that out.
I’ve rolled back to 0.91.4 to see how that goes as it has been very stable (Intel NUC).
I didn’t have errors either, but in my log were some messages about components being installed. The last one wasn’t finishing, so I stopped home assistant and installed it manually (I’m running on hassbian, so installed in venv). After that, all is good.
You might look through the logs and see if HA is trying to install something on startup.
Home Assistant is a good open source software, but it is still far from a stable release.
The development is focused on the devices integration but the software architecture is incomplete. To reach a stable release, it is necessary to build a strong software architecture and to address: the component integration strategy, the security, the front-end UI for the devices customisation and the upgrade strategy. At the moment each upgrade is destructive, after the software upgrades, we need to review customised components, device customisation and front-end cards.
Am I the only person who has a complete working system after an update? Am I just lucky? Am I part of a silent majority who have stable systems?
Yes I have been using Linux since the late 90s. Yes I can type at a command line (but if you can type in a forum post, you can type at a command line). In other words I don’t think I have any special skills.
Is my system faulty? It just seems to work. I am concerned that I am alone here.
Same here, had a few problems earlier, now I read the release notes, assess if I am impacted and if I need that update. Then I usually wait for the initial storm to quieten down.
By then I have a good idea what will break and how to resolve it.
Running docker on unRaid
Backup on github
Of course I understand that I use the community as my Guinea pigs. And I am thankful for the brave people that test the beta version as well as the early adopters of every release.
Seems my issue might possibly be related to a script I recently added?
- alias: 'Check if alarm radio flips to idle'
initial_state: true
hide_entity: true
trigger:
platform: state
entity_id: media_player.ensuite_speaker
to: 'idle'
condition:
- condition: state
entity_id: alarm_control_panel.home_alarm
state: 'armed_away'
action:
- service: script.turn_on
entity_id: script.radio_alarm
If this script runs, it seems to kill HA and the front end can’t be accessed. Only a host reboot resolves the issue. Maybe I need to stop the script before I start it again?
I’ve had an issue similar to what you describe for a few weeks. First, automations stop working and uptimerobot informs me it cannot reach my homeassistant. The frontend is not reachable but for a while at least I can access samba, so I check the log and there’s nothing in there. Hassio can be stable for anywhere between 4 hours to 5 days; previously it has been stable for weeks at a time with no hardware power cycles or reboots.
After a LOT of trial and error I’ve concluded that it started when I added the stream: option to my config. I have two dafang cameras, one of which has only about 95% uptime due to its positioning at the outer range of my wifi network. Various posts in this community have suggested that HASSIO can hang if you have stream: enabled and the connection to one of the cameras is not completely stable. I did notice that some of these freezing occurrences happened shortly after I had remotely checked the camera streams. And I don’t recall this issue happening before I happily added the stream: option to my config.
Having removed stream: from my config HASSIO has been stable for 3 days, even with me checking the cams every now and then. I miss the ability to stream the cam feeds to my TVs but stability is more important. Hopefully this is definitely the issue, because I thought it was my tinkerboard, so I moved to a VM on my server, and then back again when the issue didn’t resolve itself! Then I thought it was an MQTT issue so I spent considerable time making configuration changes to MQTT devices and the broker.
This reminds me of a similar issue last year when binary_sensors were killing hassio every 12 to 48 hours, very similar to this issue. There was nothing in the log that could point me in the right direction until I decided to peruse the github Issues section for home assistant and discovered the binary_sensors issue. There didn’t even seem to be much discussion about it in the community forum.
I have a high tolerance for issues like this in beta open source software but having such serious failures, with no indication in the log, can make life with home assistant very stressful. Stability needs to be absolutely paramount with a platform that is basically running your house. Don’t get me wrong, home assistant is great and the developers / community owe me nothing, but there have been a few times when I’ve considered scrapping the entire thing out of sheer frustration.
Thanks for the feedback. There is a conflict between the stream component and shell commands (which I use for PTZ control) noted here curl shell command hangs HA · Issue #1068 · home-assistant/supervisor · GitHub. I’m just testing rest commands now as an alternative which is a comment made in that issue thread.
I read about the shell_commands issue; I use quite a few rest commands to communicate with the camera PVR software but not shell_commands.
I can’t remember where I read about the unstable camera network connection causing the stream component issue, but there are other similar reports with circumstantial evidence that the stream component is locking hassio:
(or rather, what do you believe the issue is, and why would fixing his wifi fix it?)
In my case I’ve ha HA lock up a couple of time since moving to 0.92.2 (as seen above) - I dont believe there is any issue with my wifi network (or more specifically, no change in the last few months)
