Hi HA-specialOPs.
I use the latest HassIO(9.5) with latest HA(2023.2.2) at a VM on an dedicated unraid server (run for >1year without issues).
Some days ago my hassio started to have weird connection issues:
- some lovelace tabs are not show content or miss some entitys
- refresh of lovelace contents don’t work or not shown (images, maps, icons)
- sensor data is missing
- HACS package updating load forever
But:
- every URL that timedout in the log works fine with a curl request from the cli on the Hassio box
- some of the issues are gone if I open a new Tab with HA, but reload the tab don’t work
In my home-assistant.log I have kind of these messages:
homeassistant.components.sonos.exception.SonosUpdateError: Error calling SonosMediaPlayerEntity.media_play on media_player.sonos: HTTPConnectionPool(host='10.X.X.X', port=1400): Read timed out. (read timeout=9.5)
WARNING (MainThread) [custom_components.hacs] A timeout of 60! seconds was encountered while downloading https://raw.githubusercontent.com/ProudElm/solaredgeoptimizers/v1.2.0/custom_components/solaredgeoptimizers/__init__.py, using over 60 seconds to download a single file is not normal. This is not a problem with HACS but how your host communicates with GitHub. Retrying up to 5 times to mask/hide your host/network problems to stop the flow of issues opened about it. Tries left 0
ERROR (MainThread) [homeassistant.components.websocket_api.http.connection] [139690846196960] Error handling message: Unknown error (unknown_error) from 10.X.X.X (Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/106.0.0.0 Safari/537.36)
WARNING (MainThread) [homeassistant.components.calendar] Updating caldav calendar took longer than the scheduled update interval 0:01:00
WARNING (MainThread) [homeassistant.components.websocket_api.http.connection] [139690888959104] Disconnected: Did not receive auth message within 10 seconds
Disconnected: Did not receive auth message within 10 seconds
Console in DeveloperTools shows something like this after a minute or so, but requesting the timeouted URL with curl works:
Restart hassio worked for some hours, but after 3 times of “problem->restart->problem” cycle, this seemed not a temporary issue.
I think network issues are excludet, because a curl request of every timeout url work from the same Source and open another tab shows images that are not shown in a broken tab.
Also a Problem with a plugin seems unlikely, because of the core problems of all kind of connections from HA.
- CPU/RAM/LoadAverage are all good
- Disk hav ~3GB free space
- Observer page shows: (what das this mean)
|Supervisor:|Connected|
|—|—|
|Supported:|Unsupported|
|Healthy:|Unhealthy| - hassio->system looks good, but 85% disk usage: (will try to resize the disk, but this should not lead to connection issues) - sorry, but only one image are allowed for new users
Hope anyone can help me where I can found more information or how I can debug.