Actually I am unsure whether its that or the os-agent that needs updating.
for me it worked after both. First updated agent to 1.5.1 and then rerun install supervised. Now network storage is compatible.
But my sound still does not work… 
rhasspy also shuts up 
Did 2023.6 update kill HA for anyone else?
I have Raspberry Pi 3B+.
Performing the update it does not boot up anymore.
I luckily have SSH enabled so I can still SSH into the device, but HA is dead.
Wonderful update feature-wise, but for some reason my HA instance became a lot less responsive after installing it! I notice screens in a loading state that I absolutely never saw before. Sometimes I find myself hard refreshing the page because it takes so long (it doesn’t help). After a while the data comes through, but it just feels sluggish.
Anyone else experiencing this?
This is on a Tinkerboard with data stored on MariaDB (on a Synology NAS).
Check your cpu and memory stats
If you have high cpu usage, capture a callgrind file with the profiler.start service (see the profiler integration docs). Also capture a py-spy recording Instructions to install Py-spy on HAOS
If you have high memory usage you can use the profiler.start_log_objects service to see if something is leaking python objects.
And now for a wild guess: The current images include python 3.11.3 which has some problems leaking ssl objects with some web servers (won’t happen to everyone). cpython 3.11.4 fixes the issue but the docker images were built before it was released so it wasn’t available at the time of the release
3 Likes
I run HA on a Pi4 with HAOS and supervised on a Pi 3B+. The Pi 3B+ updated but I found that Supervisor was not working after. Rebooting the Pi made no difference, I managed to rollback to 2023.5.4 as I still had access to HA.
Can you run the rollback via SSH ?
Regarding the new “Integrations” page:
a small observation about these “custom” & “cloud” badges.
2 Likes
No problem here updating my RPi 3B+ to HAOS 2023.6.1. I know that doesn’t help solve your problem, but at least you know it is possible. With SSH at least the HA Core commands are available.
Anyone know if it’s possible to reference the “share” mount point in configuration.yaml?
I see something like this has been asked before:
I mounted a share as “HALogFiles” and it shows up in the file editor in HA as a subdirectory under the directory called “share” which is at the same level as “config.”
I’d like to define a file in this directory as the destination for a notify: in configuration.yaml. I started with something like this:
notify:
- platform: file
name: testlogfile
filename: share/HALogFile/testlogfile.txt
I get an error because HA prefixes it with “config” to become
/config/share/HALogFiles/testlogfile.txt
Is there a way to tell HA to look in share instead of config?
/share/, not share/
2 Likes
That did it, thanks!!
Thanks, just tested it and indeed works. Still have to tweak it to send it to 1 Nest and not all. But you probably can’t delete the accounement text.
EDIT:
If your grocy is having issues starting, you may have to update to beta version 4.9 onward temporarily.
Thank you @reste_narquois for the solution in next comment
==================
It looks like some integrations that are purely entity based (do not have any devices) have no way to restart the integration through the UI like before (at the integration level) if they had severe startup errors. (grocy is notorious for loading the integration prior to the grocy server itself, throwing python errors) normally a reload after boot resolves it.
Seems to only be affecting my grocy, has anyone noticed similar issues with other integrations now that reload as been moved inside the integration page?
You are not seeing the reload option there because your Grocy component hasn’t even loaded in the first place.
Check your logs to see the error message. There’s an issue reported for the component, which also includes a work around to get it running again
1 Like
Great feature with the Network Storage option so we can skip Google Drive backup.
But shouldn’t Samba/windows protocol not be the most simple way to connect? I can map my NAS share to my local PC but I cannot map from my Home Assistant. What do I do wrong? (It should be so simple):
Yes, this is like making a mistake in a yaml configuration. MQTT Discovery is the “automated” yaml configuration for MQTT. It has to be correct for the resulting sensor to be added. It also determines the behavior of the entity. So mistakes in it can and will cause issues in home assistant. You shouldn’t be tinkering with discovery topics without fully understanding the impact it has, and you should stick with the basic Z2M discovery setups that it provides.
Since I updated to 2023.6.1 Node Red is connecting and closing connection to http://supervisor/core in a loop so flows cannot be triggered. Restore to 2023.5 runs fine. Any ideas?
1 Like
Daikin also broken for me. Had downgrade before I had time to get some logs unfortunately but I’ll go again soon
The default Sun integration does not show on the default dashboard anymore. Is that intended?