Only screening and parsing the whole release notes took me almost 2 hours in two separate sessions. Now I got a monster list of things to check, test, change, fix, adopt, optimize, document, remove, ā¦
What a HUGE package this 2022.4 is with so many changes. I think including the update planning the whole switch from 2022.3 to 2022.4 will take me approx. 4 to 6 hours. Crazy, isnāt it Iām hoping for more changes-poor releases the next months (because HA administration is not my full-time job) haha
I hear ya! This is the first release for which Iāve actually been developing a written migration plan. All the things to record and back up before, all the steps to be ready to back out, all the existing functionality to test afterward, all of the changes to verify, and of course all the fun new features to check out after the update is done.
I just dabble around the edges, and HA administration does feel like a full-time job sometimes. Iām in awe of those who are able to really dig deep and do so much more than I.
I didnāt have to change anything with this release. Where did you have to do so, the both of you? Regarding 6 hours or migration plan, etc.? Iām really interested to hear about.
Itās all about risk. For five people, a game of Russian Roulette is no problem. I just donāt want to be that sixth player.
One: Massive database changes. This is good. Iāve been involved in discussions about database size, and committed to doing some before and after analysis of this version update. Two: I depend heavily on ZHA, which has had lots of reports of problems.There have been some significant changes hinted at, and at least one was backed out because of problems it introduced. Again, lots of before-and-after testing, and a more robust backup and recovery plan than usual. There have been reports of memory leaks and increased processor utilization. More monitoring.
Those are just my most pressing issues. For other people it will be other things. Then there are all the āusualā low-level issues which pop up with every version. Iām at the point where it makes sense to write this all down so I donāt miss a step.
After the log file config, and testing throughout the weekend, it seems that it just takes longer to connect to the gateway after a restart. With 2022.4.4 HA connects right away with my KNX gateway, on 2022.4.5 & 2022.4.6 it takes quite some time.
This is also stated in the log files: āConfig entry āAutomaticā for knx integration not ready yet: No Gateways found; Retrying in backgroundā
Does this mean that the setting āAutomaticā was removed?
So is there any way to speed things up, eg. provide the fixed IP of my gateway somewhere?
I have not open up an issue on GitHub simple because these seems more like a feature than a bug. Please correct me if I am wrong
No, it just means the discovery didnāt work - so for whatever reason no response from a KNX interface was received and processed in a given timeframe (2-3 seconds - not sure). It will just abort setting up KNX, wait for a while and try setup doing exactly the same discovery function again then.
Yes, this has always been an option. Either when configuring the integration, or by changing the integration options from the integrations menu.
There is a known bug currently when you have a KNX IP Secure interface and a non-secure interface and use āautomaticā. When the secure-interface answers first the connection will not succeed.
If it worked fine before, but didnāt after an update - and there was no mention of anything related in the relaese notes - Iād consider it a bug. Especially if it happens in a patch release.
I agree. No issues with ZHA. I have a Phoscon Conbee II plugged into an extension on my USB2 in my RPi 4 and it has preformed flawlessly for me (with Zwave plugged directly in USB3 port). I only used the extension because it was a tight fit with the zwave stick already there, and it just helps with USB3 interference. Itās only 18-24" away atm.
I run automations continuously for my Hue lights that create dynamic groups based upon inputs from the dashboard and that applies scenes of those dynamic groups on the fly. Works like a charm.
Agreed! I just took a snapshot, updated, and forgot about it. No issues. My status page is designed to alert me pretty readily if anything runs afoul in an upgrade. The biggest upgrade issue I ever had with HA was moving from OZW to zwave-js, although that turned out not to be troublesome, just a lot of time-consuming manual work renaming entities.
Media browser is cool and all but useless for most people if we canāt mount network shares for the media browser to have access to our media servers. ĀÆ\_(ć)_/ĀÆ
Ah o.k. I see/understand. I do it indeed different in such cases. I think and decide in such cases: Why e.g. changing to SwitchasX, if this does not more as my up and runing light-templates before. So I will not change.
I had brought this up a long time ago and there are Feature Request threads about it. Now that the media browser lets you link DLNA servers its not an issue for me anymore since my NAS runs the server for HA to connect to.
Not being able to queue up songs in HA is a massive oversight thoughā¦
