I also agree about the entity names and had a similar problem. I borked my whole z-wave system by using the HA editor to change the entity name of a couple of my z-wave modules. I forgot the “little” warning and used a dash in the new entity name. Can’t remember exactly how I got it fixed, but I think it had to do with some under-the-hood editing of the zwave xml file.
Don’t give up on zwave for the hardware layer. I have been using automation since my first X10 unit in 1983 and have found zwave to be the best so far; nowhere near great, but best so far.
The Aeotec Z-Stick seems to work ok for me, but it is so limited on its interfacing that it makes troubleshooting a royal pain.
Z-wave needs work, open-zwave seems to work just fine but some features can use a tweak. If you add a device and then rename it, the first part of the device is still the old name in the registry file. There are some threads abou that too. At some point i needed to make it work, since i have 70+ devices. Just 1 bad node in your network and it is screwed. Actually greated a complete post about my issues which perhaps can help others.
That said, it takes a long time to boot up (which is normal) but even with a lot of devices you should still be able to ‘switch’ a device although with longer wait times. If that is not the case then something is bad too.
As some suggested, the whole entity things changed and a good chance thats the case too. Adding a Vera only moves your problem somewhere else, and by problem i mean having another device. I use automation to also safe money on electricity by turning off the subwoofer when the stereo goes off. Adding another device consumes power too.
Room for improvement, sure! Does it work? Yup! Does it drive you crazy to the point you want to get a Vera too? Jup. But i just didn’t want to give in and it works just fine now.
I found moving from a rPI to a NUC has drastically helped startup times for zwave as well. I don’t understand how as in theory it’s radio speed limited to enumerate all the nodes but it made a huge difference.
Also setting logging levels to a reasonable level in OZW helps.
A couple dead nodes on the network can be handled by ignoring them; I do this for all my holiday outdoor zwave switches for lights. I just uncomment them when I set the lights back up.
I did move to a NUC too actually, RPI was at 100% when starting with z-wave. It saves some time but in the end it would only save a minute or 2 and then work as normal. (but now i can run more stuff on it)
Personally, similar story; currently have about 70 Zwave devices (mainly Evotronic Spirit TRVs, Greenwave Powernode 1s and three or four Fabiro devices)
I had a Vera Plus and was happy to remove it, due a lot of compatibility issues, and a total pain to resolve. Earlier comments about installing apps on the Vera as spot on, do not do it! The best things it had to offer was the extended UI to see the Network Map but in the end it was slow (many seconds) to simply toggle a socket power.
I tested moving first to open Zwave on a PI with Domoticz (as it also supported Evohome without cloud) and using a Zwave.me module This was a major improvement in a stable and usable system.
After a number of months with Evohome that had to go as it has no mesh support and it’s range was not suitable for the house (6000 sq ft), so after some more testing of TRVs I have replaced over 30 heads with spirits.
Not needing the Domoticz software I moved the core on the pi to Home Assistant, and quickly saw the pi was under strain, so moved production to a NUC and at this point I am experiencing the most reliable Zwave solution so far.
I run the front end experience with IRidium as I Do professional installations with this platform but at home I use MQTT to bridge the worlds.
Home assistant is an amazing solution and kudos to was is one of the most vibrant communities I have experienced in 20 years of Home Automation.
Cheers
Have that installed. But just yesterday wrote a macro that gets the neighbors from the Zwave states as exposed in Home assistant and this week I am hopefully to figure out how to create a custom card to illustrate the results.
A lot of valid reasons here, that’s why I went the opposite way: I split my zwave network in 3 pieces (now moving to 4). Handling 15-20 devices per network is very fast and startup is around 20-40 seconds (except battery devices). It costs a bit more but its so much easier to manage… Maybe I’ll have 5 networks in the future :). Also… I have no hops in the network, everything is direct communication
Interesting. Sorry for the rudimentary question, but what do you mean by “split” exactly? You have multiple controllers?
This is why I always read the notes and make the necessary changes if any to my configuration. Otherwise you end up 10 releases back and it’s a nightmare to get current when you want that new feature.
2 Likes
You really expect the user base to keep up with a blog in order to keep Home Assistant from breaking? That seems ridiculous to me. If I had to do that with every piece of technology in my life, I would never have any time to do anything.
I can see saying “You have to keep up with the blogs if you want to know about the latest and greatest features of home assistant”, that makes sense. Saying people should forever keep up with blogs to maintain basic functionality or to compensate for terrible documentation seems like a cop-out to me, especially for the user base that I would expect HassIO to be intended for.
You only have to keep up with the blog if you update hass to a new version. Consider it release notes.
Just like if you upgraded Word or Excel.
Also you are in an open source project that has not even made version 1.0 yet. if you don’t want to learn and help development, then go for a product where you can legitimately grumble if it goes wrong, because, y’know, google, amazon et al are great at fixing stuff that doesn’t work, and warning users of breaking changes.
2 Likes
That is not what anyone said. For the time being yes you do need to keep up with the breaking changes. Home Assistant is not at the point where things will be super easy. As you have seen throughout all of the updates coming in it has been getting better. Look at how easy it is configure Hue after the past 2 releases. It is getting there but there will be breaking changes along the way. All users are saying is that you need to keep this as part of the upgrade process until things are where they should be.
Developers try their best to avoid breaking changes but in order to move forward it has to be done. Look at the python upgrade, it came time for it to get done. Eventually yes things will get better but it will not happen over night or in a single release, major changes take time. I do agree that is not ideal but if you got this far in home assistant then you understand the process of things. You get how to create an automation or a group, just make this part of the upgrade process and it will be pain free.
I definitely have a LOVE-HATE relationship with Home Assistant. I am an electrical engineer (circuit design), so I have some basic experience with Python, mostly to write procedural scripts to automate data collection. I started using Home Assistant in November 2017 and I have only used HassIO since I know next to nothing about Linux or computer networking.
Initially things were very rough when I tried running Home Assistant from my Windows machine. Holy cow, that was a real nightmare. I never even got to writing any automations, nothing worked. I gave up and got an Raspberry Pi 3, the first time I had heard of the device. Things were quickly much, much better.
The last additions to my home automation system have been Schlage ZWave locks, Dome ZWave doors/window sensors and an Aeotec Multisensor 6. Three times in a row now I have been met with problems that consume hours of time to get the devices integrated and functioning in Home Assistant.
The recurring issue I have is that I constantly feel like I am looking for an instruction manual that doesn’t exist. I would be willing to pitch in and help with some documentation of things I have used, but I don’t even know how I would go about doing that. I assume trying to figure out how to help with documentation would just lead to looking for another manual that doesn’t exist.
I guess I am not sure who Home Assistant is intended for. Is it only supposed to be for computer programmers or for the average person?
If you go to any page of the official documentation you will see a “Edit this page in Github” link near the top right hand corner.
Click on that and edit the page in github to provide updates. If everyone did that every time they added a device and updated with their findings the docs would be a lot better but that doesn’t happen often. We all try to update them often but we don’t all use the same components. You should stick to the home assistant documentation as much as possible (like read it more than once and follow it to a T) and when you have a doubt hop on in Discord and someone will be able to help you get sorted. Lots of good people out here who want to help users and make sure the docs are up to date.
1 Like
3 sticks in 3 pi’s, shared over the network with ser2net, 3 home assistants that only do zwave and connect to zwave stick via socat.
The 3 home assistant instances publish state and take commands via mqtt (a mosquitto instance). A 4th instance gathers & controll everything via mqtt event stream, mqtt state stream and some custom automations.
Basically I treate the “slave” instances as drivers - once I get everything set up and configured, I freeze them and not upgrade until I need a new feature for those devices.
Another “slave” integrates with broadlink ir and so on… I’ve got around 10 instantes on different version (which i found were rock-solid).
If I bough zigbee devices or ambilight or philips hue, I would add another home assistant to it.
I’ve had “broken” versions for various things but I just roll back one of the slaves, the rest of the automations work fine.
I’ve switched recently to docker from proxmox and the whole stack (with mysql, influx and others) barely uses 2GB of ram and the cpu is around 5-10% with very small spikes on a 10 watt 4-core processor ( http://www.asrock.com/mb/Intel/Q1900B-ITX/ ). Now to upgrade i just modify the version in the docker file from v0.6x to v0.6y, run a command, then put 24 in a another file. After 24 hours a script will run and do a git reset --hard and re-deploy the old version unless I commit (save) the configuration with the new version. I can put any number of hours to make the whole stack revert to whatever is last saved.
Instead of pi’s + sticks (which can be quite expensive added up) you can use a vera (very stable with the vera software disabled, but not so much with the default integration) or, the cheapest i went for a friend as a bunch of mr3020 with openwrt https://openwrt.org/toh/tp-link/tl-mr3020?s[]=mr3020 . I use the pi’s for bluetooth detection and I plan to also add cameras to them for presence detection (i’m learning machine learning now 
)
I’ve automated most of my stuff in docker, so i don’t really need to watch it (except for hardware breaking). For example a container whatches the mysql containers - it backs it up every X minutes and if it breaks it re-creates it then uploads the last backup. I’m now working on an auto-update docker to watch home assistant, automatically do the update after version release + 2 weeks, watch it for issues (machine learning agan), then downgrade/save it based on results…
I think i wrote too much… I could tak all day about this… I mainly wanted to avoid the pain of v1 works for A but not for B, v2 works for B but not for C and so on… this way each “driver” does it’s on thing… if that breaks i only have to fix 1/10th of the home automation not all of it…
Applying this to the z-wave lock leads me here: https://github.com/home-assistant/home-assistant.github.io/blob/current/source/_components/lock.zwave.markdown
Am I supposed to go to the Issues tab at this point? That tab doesn’t seem to be unique to the z-wave lock page and once I click the Issues tab, if I then click the the <>Code tab (which is where I started) it doesn’t go back to the z-wave lock page. I don’t see where I would report anything on the <>Code page which is specific to the z-wave locks.
If you are submitting a bug on the documentation then yes you go to Issues and submit a bug about what page the issue is on.
If you want to submit your own changes to help out the next person, there is a icon that look like a pen on that page click on that and from there you can apply edits. For formatting you need to refer to githubs mark up:
https://help.github.com/articles/basic-writing-and-formatting-syntax/
After you make your edits you submit a Pull Request and read the pull request template and fill in the requested info (i.e. why you are submitting the changes). Once submitted someone will review it and it goes from there. I may be a missing a step but thats the overall flow of it.
Sorry to bust in on the Confo… I am a Newbee here and i was just wondering abut why do most folks hook the HA system to the Vera?? Is the Vera users interface(GUI) that bad? I am useing a RPi3B+ now with HA and USB stick… I also want to beable to run a WIFi temostat EMERSON type with Z-wave switches and dimmers. I think I need a VERA like HUB to do this?? Correct.
No you can configure all of that with the usb zwave stick and there’s probably a component for emerson. People moving to the VERA hub just don’t want to deal with the home-assistant zwave interface. It and openzwave can be a pain in the ass. using the VERA hub bypasses that.