New Supervisor Version 139 broke my Hassio (Again)

Installation Home Assistant OS
1

The last Supervisor Update except 138 have broke everytime my hassio.
Today i update the 139 Version. Since that Time i can not load the hassio tab, again.
Every Sensor is not working and i know if i reboot the system or make it powerless wont go online again, like the last few times.
I can not restart hassio. ssh or webssh it shows the error:

The http Request failed with the error: post http://hassio/homeassistant/restat: dial tcp 172.30.32.2:80 connect:no route to host.

The hassio panel show error:

Unable to load the panel source: /api/hassio/app/entrypoint.js.

can i downgrade the supervisor version to 138?
the last few times i had to reinstall the system, but with the actual supervisor version it takes few day till the error cames again. i do not have any mqtt devices configured.

Hassio is actual useless and not really online.

Has somebody the same Issues or better, an Solution for that?

Thx

1 Like
2

So updating the supervisor always breaks hassio for you…so you updated the supervisor to a brand new version? (doesn’t even show up for me yet as an option)

No sure if its supported for the supervisor but run the update command via ssh and specify 138 as the version.

hassio supervisor update --options version=138
3

Now the system crashes and it wont come up anymore. i can not do this.
other options?

4

Without any ways to debug the system its hard to debug it.

Last ditch effort is to try and mount the disk and pull the log files using another computer.

If not, I hope you have a back up.

5

I get the same thing happening on my system. I thought I was going crazy… Everything was perfect (in the process of setting up my new home with new Hassio system – hasn’t been without problems to say the least ) but now this happens… very frustrating … Been learning a lot about docker and linux though at least! I don’t like to post support issues about anything or any product unless I have exhausted all other options and resources… but this one, I’m, at a loss… Hassio tab won’t load. most addons just disappeared from my config… (the containers aren’t even running) and I can’t load configurator, even though that container is running… Hassio tab shows that odd entrypoint.js error (which has the output of nothing in the iframe - kinda looks like its trying to load, but its not). And every time I reboot (I’ve rebooted my Nuc about 50x past 2 days) or restart (at least 150x) something different is happening… I don’t even have too much loaded on the system yet… 4 integrations, only about 10% of my devices… oh and the Zwave stick (aeotek gen 5) isn’t getting recognized anymore about 50% of the time. I’ve combed through probably 150 million lines of logs… googled every single error message… and well. I’m here posting this so it didn’t work… I’ll try updating via cmdline to earlier version… I dont know if its hassio_supervisor causing this or not but the OP seems to think so - and I can’t do much more damage, so I’ll give it a go and post the results…

What is so strange is that the interface loads, but its like its a dummy interface… most of the time, nothing that does show up on the interface, won’t react to any triggers, like turning a light on or off, etc… sometimes service calls work, most of the time they don’t… only about 30% of devices load… and the logging? either it shows absolutely nothing wrong, nothing in the logs… or they are getting hammered and I couldn’t google errors fast enough…

6

i have an backup, but it takes much time to set up everything working to see few days later the same issue.
it all starts few supervisor versions before since 138. with 138 everything works fine.

where can i found the log files?
i have downloaded linux file systems from paragon an have full access to the docker image right now.

7

Perhaps this can help

Delete the discovery.json file in the hassio directory (/usr/share/hassio in a standard docker installation of hassio. Not sure about hassos)

8

Thx
They had problem with the 138 and they update after the issue to 139. i have a problem with the 139, 138 worked fine.
I can not found the discovery.json. I use hassos. Where is the file located in hassos?

9

The problem seems to be related to the mqtt add-on, more than your hassio update. It might just be a coincidence. I don’t have an hassos running right now… Keep an eye on the GitHub issue, I’m sure you are not the only one with this question.

10

I don’t even have Mqtt addon installed. but thank you, I’ll check out the issue thread.

11

I have a duel boot Win10 PC and Kunbuntu. Running from Kubuntu there is a drive called Hassos and a folder called Supervisor and discovery.json is in there. I had to delete it using terminal “sudo rm discovery.json” to remove the file and Hassos is back up and running for me.

I am still trying to sort out this MQTT issue in the meantime but at least I have SSH and Samba access back after being down all night.

12

Confirmed - deleting discovery.json fixes the issue - though i was on build 139 already

13

Is anyone else still having trouble even loading homeassistant like I am? I never installed MQTT so thats not the issue… HA just won’t load… it says initializing or something… I’ve tried every trick mentioned on any of the 30 threads complaining about this last update (is there really no word from the developers yet about what happened, how to get back online, anything?)…

I’ve rebooted, updated everything, on the new supervisor,… I have no discovery.json file anywhere - two things happen… Eithert after restarting it has no errors but just won’t let me in the interface (nothing else works either) - or it goes into a booting up loop with supervisor… basically it looks like the internal network never comes up - it can’t shut down various containers… reloads containers., but file systems can’t be accessed, all sorts of stuff… and it just keeps recycycling supervisor - trying to load… says hassio is running but nothing works… Been down 2 days now… Can’t get back in.

I have file system access (i run this on docker/nuc setup) and I even have backups… But I have no idea how to restore this properly… because I cant’ access the snap shot interface. Can I do it via cmdline? and if so, what are then the steps to update properly, or what don’t I update, do I comment discovery on my config file, etc etc…

Thank you!

14

ĂŤll try it and give a Feedback, thx.

15

discovery.json not in the Folder. only addons.json, config.json, homeassistant.json and updater.json.
i think its an other issue than mqtt. when i plugin the monitor to hassos and it boots up, it says not network connection found. the leds on nic not blinking and it stucks on start kernel. :frowning:

16

Yep - same here…

17

This is a major problem with this forum, useless stuff like this. Either report the error messages properly, or don’t bother at all.

1 Like
18

You’re right, I apologize. It was really late (or early) and had been staring at my screen for too many hours. It wasn’t an error message I suppose. The only thing I get when trying to look at the UI is “Initializing” below the blue header row of Home Assistant. Nothing else shows for the front end.

In Homeassistant log (running Hassio) - it doesn’t have anything besides SSL errors repeated like:

2018-11-06 13:46:26 ERROR (MainThread) [homeassistant.components.updater] Got unexpected response: None
2018-11-06 15:38:51 ERROR (MainThread) [homeassistant.core] Error doing job: SSL handshake failed
Traceback (most recent call last):
  File "uvloop/handles/stream.pyx", line 609, in uvloop.loop.UVStream._on_eof
  File "uvloop/sslproto.pyx", line 171, in uvloop.loop._SSLPipe.feed_ssldata
  File "/usr/local/lib/python3.6/ssl.py", line 689, in do_handshake
    self._sslobj.do_handshake()
ssl.SSLError: [SSL: UNKNOWN_PROTOCOL] unknown protocol (_ssl.c:841)
2018-11-06 15:38:51 ERROR (MainThread) [homeassistant.core] Error doing job: SSL error errno:1 reason: UNKNOWN_PROTOCOL
Traceback (most recent call last):
  File "uvloop/sslproto.pyx", line 504, in uvloop.loop.SSLProtocol.data_received
  File "uvloop/sslproto.pyx", line 204, in uvloop.loop._SSLPipe.feed_ssldata
  File "uvloop/sslproto.pyx", line 171, in uvloop.loop._SSLPipe.feed_ssldata
  File "/usr/local/lib/python3.6/ssl.py", line 689, in do_handshake
    self._sslobj.do_handshake()
ssl.SSLError: [SSL: UNKNOWN_PROTOCOL] unknown protocol (_ssl.c:841)

In my host (ubuntu) syslog - it complains of runc version - repeatedly, here’s an example:

Nov 7 13:06:55 hassio-nuc docker.dockerd[2627]: time=“2018-11-07T13:06:55.492571872-08:00” level=debug msg=“sending heartbeat to manager {2hyfd14wbu5lu2tf066v8h0lq 192.168.1.126:2377} with timeout 5s” method=“(*session).heartbeat” module=node/agent node.id=2hyfd14wbu5lu2tf066v8h0lq session.id=icqr3xbyww0rd450tqofxxkqw sessionID=icqr3xbyww0rd450tqofxxkqw

Nov 7 13:06:55 hassio-nuc docker.dockerd[2627]: time=“2018-11-07T13:06:55.493382453-08:00” level=debug msg=“received heartbeat from worker {[swarm-manager] rwupmf0bekiz1fvel35d9neqe 2hyfd14wbu5lu2tf066v8h0lq <nil> 192.168.1.126:41344}, expect next heartbeat in 5.122414s” method=“(*Dispatcher).Heartbeat”

Nov 7 13:06:55 hassio-nuc docker.dockerd[2627]: time=“2018-11-07T13:06:55.493907729-08:00” level=debug msg=“heartbeat successful to manager {2hyfd14wbu5lu2tf066v8h0lq 192.168.1.126:2377}, next heartbeat period: 5.122414s” method=“(*session).heartbeat” module=node/agent node.id=2hyfd14wbu5lu2tf066v8h0lq session.id=icqr3xbyww0rd450tqofxxkqw sessionID=icqr3xbyww0rd450tqofxxkqw

Nov 7 13:06:56 hassio-nuc docker.dockerd[2627]: time=“2018-11-07T13:06:56.101237419-08:00” level=warning msg=“failed to retrieve docker-runc version: unknown output format: runc version 1.0.0-rc5+dev\nspec: 1.0.0\n”

I’ve searched for ways to recover, and have asked in various places but most ppl had issues with Mqtt - however I do not use that (yet).

Thank you for bringing that to my attention though. I’ll be more clear and specific with any future posts.

1 Like
19

This is a major problem with this forum, useless stuff like this. Either offer to help, or don’t bother at all (there’s already enough dicks in the world).

Funny how that works both ways, huh?

20

Really you are way off beam. I do at least know that this is a technical forum which requires questions/problems to have technical information. As a result of my post, technical info was supplied, and now someone may be able to help. My post received a like, and laresentim provided some helpful info.

You, on the other hand, have provided nothing to the conversation.

This forum could do with some moderation IMHO.