Well I would imagine so its just the one that comes from the docker community site I believe but will check and report if different.
Mine is Intel so thats X86 I thought?
Well I would imagine so its just the one that comes from the docker community site I believe but will check and report if different.
Mine is Intel so thats X86 I thought?
X86 = 32 bit. AMD64 = 64 bit, whether it is Intel or amd.
haha…cheers for clearing that up…as you can see I’m not really up with ‘architecture’ type stuff.
Seiously though has anything been found out as to why the ‘memory leak’ started and when we can expect it to be sorted? I don’t get on discord nor do I monitor the github issues as I just don’t have any extra time atm.
No, but we can help I guess (remember I do not have the problem)
Enter your container, install mem_top, run
docker exec -it home-assistant /bin/bash
# You are now in your container
pip install mem_top
python3
#the python prompt will appear
from mem_top import mem_top
print(mem_top())
You will get stats, they mean little to me, but as it says on the pypi page,
Notice which counters keep increasing over time - they are the suspects.
Okie…I’ll have a crack!
Using 102.2…When I use the GUI to restart the server, the logs are revealing a complaint about unable to install AV==6.1.2. I do not have “stream” configured, nor do I have cameras with ffmpeg platform (I do use platform generic and local_file). I don’t recall having this issue before…any clues as to which other component may be requiring this??
2019-12-11 10:17:30 ERROR (SyncWorker_5) [homeassistant.util.package] Unable to install package av==6.1.2: ERROR: Command errored out with exit status 1:
......
2019-12-11 10:17:37 ERROR (MainThread) [homeassistant.components.homeassistant] Component error: camera - Requirements for stream not found: ['av==6.1.2'].
If you don’t use the stream component you can safely ignore the error.
So I finally got around to messing with this more. I exposed scripts to Alexa and they wouldn’t show up.
Not sure if this issue was just for me or if there is something specific in the update that caused this, but I had my Alexa settings configured in the my configuration file. I had to hashtag all that stuff out, then configure it via the GUI and then everything popped up in my Alexa app as it should. Not sure if this is an issue that needs to be looked at, but just a heads up for anyone else that comes across this.
I had to hashtag all that stuff out, then configure it via the GUI
If you’re using cloud I think that is the way it’s supposed to be.
I can’t remember if I got back to you after our conversation about this a couple weeks ago but I did end up setting up my Google Assistant config manually. New entities are no longer being exposed automatically as desired. I still have access to the Cloud UI interface for Google Assistant but I cannot toggle any entities on or off (which is just fine by me) but I can see which ones I have exposed via the config.
Just to give an update where I am at now @nickrout. I didn’t have time to try the stats thing you posted above. Instead I have been chasing my tail and been through all the releases up to and now including 0.103.3 and can now report that my CPU and Memory stats are looking pretty good AND most importantly after about 5 hours of runtime after installing 0.103.3 I have ZERO errors in my log! HAPPY!
By now my log would have been cluttered with rubbish and CPU and Memory Stats would have been getting nasty! So Merry Christmas to all and I hope you all are finding a similar result!
PS: For future update choices can I ask if I was to just choose the ‘latest’ option when choosing an update…is that considered to be a stable version? If not how do I find out the latest ‘stable’ version?
If you pull ‘latest’ you will get beta/dev versions. Pull stable instead.
Releases here https://hub.docker.com/r/homeassistant/
@DavidFW1960 Cheers, but I’m pretty sure my container station looks at that location when I choose the ‘Docker Hub’ option when I ‘create’ a new container…
After that I get a huge list of seemingly EVERY HA RELEASE…In all that huge list I have never seen one labelled as stable…Until as I am writing this (and moving through the create process)…finally right at the very bottom is indeed ‘stable’…How I could ever have missed that is beyond me but I have honestly never seen it. I have always looked carefully through the list for the latest release as shown here because in the past I did get a very bad release by choosing ‘latest’.
So I continue on to add the required config for my install and create it. Surprise, surprise it’s 0.103.3
Can I expect that next release (0.104.0 possibly?) will the be also listed as ‘stable’?
If that’s how the system works then what if another item was added to choose the last ‘rock solid’ release…or whatever the person who set’s up these things should care to call it?
In any case cheers for your advice David and all I can say is 0.103.3 seems (with no config changes from me) like it has finally solved some ongoing problems.
Yes when 104 is released it will be tagged stable
I would not have described any release since 0.100.0 (except 0.103.3) as being stable by any stretch of the imagination.
NOTE: Nothing was changed/edited in response to breaking changes affecting me in my config in all the versions I have tried up to 0.103.3.
However I did add the fronius custom component @ 0.103.2 or 0.103.1 though.
For you maybe, but HA has been stable for me for a long time.
It also depends on whether you are using one of the supported methods of installation, remember they can’t test every edge case installation method
@aidbish Looking good now I have 0.103.3 flick the switch back to 0.103.2 and errors are back within a few hours. Same config.
Flick the switch back to 0.103.3 (Same config) about 6 hours ago and zero errors so far
FYI docker installation pulled from docker hub
That makes sense. If 0.103.2 didn’t have any problema, then 0.103.3 would never have been released.
The main releases 0.XXX.0 have new features and breaking changes. After the main releases new features and breaking changes are frozen and the patches (that last number increasing) are only to fix issues that get identified by the community.
You could say that the .0 release is unstable, but that’s not fair either, since it just doesn’t have any known issues.
Flick the switch back to 0.103.3 (Same config) about 6 hours ago and zero errors so far
Is there anything in the release notes that says what was fixed that would have affected this? Did you check the issues with 0.103.2 on Github? The ‘dot’ releases are for bug fixes and unfortunately, the dot releases aren’t tested as betas before being released whereas the .0 release has has 2 weeks of open beta testing.
Just FWIW, 0.103.4 just dropped as well… so if you have problems with that just roll back to the last one YOU find to be stable.
I’ll be installing 0.103.4 and if it performs as well as 0.103.3 I’ll continue with it (0.103.4)
FYI No errors since I flicked back to 0.103.3 yesterday…it seems like a pretty good release to me.