Current Hardware: Vera Plus, HA in Synology Docker
Available SW / HW: Multiple RPi's, Z-Way UZB w/lic, Z-Way Razberry board
ZWave Devices: 30-ish early gen GE/Jasco light switches
Issue: I've gotten myself into a spot that I'm not quite sure how to safely get out of, and I'm hoping someone on this forum can help me. I would ultimately like to get to a point where I'm using Home Assistant in Docker on my Synology as my UI for turning on/off lights, but I've read the "Pyramid" post and am open to other options. I can currently ssh to my Vera Plus, but cannot access the web UI (get a 403).
How I Got Here:
I installed HA in Docker w/socat on my Synology NAS, and things were running perfectly for a long time, but I was forced to upgrade HA in tandem with an upgrade to the iOS app. So, I chose to use a Docker build that already had socat built in.
Then, I saw that there was a new version of firmware for Vera, and decided to upgrade to Firmware v.7.0.31 around Apr'20. This was a mistake in a big way, and I sought help on the new version of the Vera forums, which eventually led me here. I also re-ran @rafale77's mod files on the newer firmware version, and I think I shot myself in the foot there. I knew it was gamble, and take full responsibility. I've learned a lot and have benefitted greatly from his work.
I tried to run the Nuke Vera script, as that sounded exactly what I needed my Vera to do - take commands from HA and nothing else. This seems to be working, but I'm unable to control my network with HA.
Additional things I've done / tried / worry about:
Validated that HA is sending and receiving responses via socat from the Vera in the HA openzwave logs. I see logs that seem to indicate that a message is sent, and a message comes back from the Vera, but lights don't turn on/off. I've clicked the Heal Network, Test, Network, and Soft Reset buttons in the HA UI too.
I purchased a razberry. Then, I ordered a UZB, after seeing this post about migrating, assuming that it was best to try to follow that guide verbatim and not have to exclude / include a bunch of devices. I'm leaning towards completely divorcing myself from Vera.
I'm not 100% sure I have a valid backup of my zwave config. There are no dongle.dump files in /etc/cmh whch may make sense, as I was using after an external drive w/extroot. I do have some in ./mnt/sda2/etc/cmh/. It looks like I've got extra, actually, as it looks like I probably made backup copies some where along the way.
I shutdown the Vera and removed the extroot drive, reversed all of the changes Nuke-Vera made to the init.d start up files, and rebooted. I did this because the migration guide mentions to set the path of the z-way UZB in the UI settings, and I couldn't find how do this over ssh. I also don't know how to backup the zwave config from the CLI and was going to do this via the UI. As mentioned previously, I'm getting a 403 "Access Denied" message. Looking at logs and the config for lighthttpd, I think it is trying to follow some symbolic links that no longer exist, but I'm hoping I can get out of the mess I've made without troubleshooting and solving the HTTPD issues. I think I just need to know how to follow the steps to clone the Vera to the UZB via the CLI?
I found a command online method that allows me to view the zwave keys with hexdump. The keys on the device are different than the keys in on the drive that i was using for extroot, and that concerns me a bit.
If you've made this far, thanks for your time! Does anyone know the best way for me to get myself out of the mess I've got myself into?
All the best,
In January 2020, I decided to get involved in automation as a complete beginner.
I purchased a Vera Plus device controller.
I built the network gradually, at the cost of overcoming no small problems .....
Since I am a beginner I have had a lot of problems.
For example: I had 6 multisensors that sometimes got stuck with motion detection (lights stayed on) and also created ghost devices.
I knocked everything out of the system, the system stabilized a lot after that. (many red lines disappeared from luaupnp ....)
Over time and with the help of the experts (Eg: AKbooer was a great help to me in netatmo which I thank again) I managed to solve most of the errors.
I only use a reactor to control the logic and execution.
The logical composition of my reactor is not very complicated in my opinion.
My DSC alarm (with plugin) controls the house modes, cameras, each room has a separate reactor sensor (18 sensors in total) which control the automatic lighting and shutters.
I tried to improve the conditions and scenes in the reactor to perfection. (don't have executions, delays and reservations at the same time, I record the states with the "time from reactor" group state, just like the motion and opening detections from another sensor.
The situation is that the system does not always work reliably.
I experience Lua recharges at completely unpredictable intervals, about 1-3 per day
Because of this, unfortunately my scenes slip, (or don't happen)
For simple things (tablet charging in) there is usually no problem.
However, if I am not at home for several hours and return home, the alarms will turn off, and the luup recharge(90%).
If you approach the afternoon when more times are set in the time reactor then recharging will still occur but not always.
For the past 10 months, I feel tired of the constant error search and still not good. I'm starting to doubt Vera's abilities.
Therefore, I would have the following questions:
Are the conditions set in the reactor loaded on the vera without action?
Can the Vera be stable and well-functioning without an unexpected lua reload?
If so, what do I need to add? (I'm a simple user, not a programmer, I don't know lua and linux programming languages and I don't want to learn because of vera)
I don’t want to deal with this too much anymore, so I still spent more on it than I deserve.
Is there another controller that works safely and well on the z-wave platform?
In what direction should I move away from the vera, which will not disappoint me?
Have you heard of a controller called comfortclick? (I was now offered this instead of vera)
Thank you in advance for your help.
So ... I am trying to do the following:
Vera Plus (1.7.4955) Zwave only -> HASSIO VM (full control via Vera Integration)
HASSIO VM (leveraging zwave.me UZB1 dongle via USB)
I have done the following so far:Updated UZB1 dongle (via RPi SmartHome) to latest 5.39 Firmware
I followed the following steps to try and migrate my existing Z-Wave NW (currently on Vera) onto the UZB1 dongle following the steps listed here:
Steps 1-3: went fine (although the dumps are labeled as dongle.6.1.dump.x)
Step 4: went fine and the UZB1 was recognized as it should be (per dmesg)
Step 5: updating the port to /dev/ttyACM0 went fine, although I didn't see any indication of luup reload (or a save button for that matter when updating the port mapping)
Step 6: I did the touch for dongle.restore, but wasn't sure where to trigger a luup reload (I assumed it was Z-Wave Settings > Advanced > Reload Engine). I believe I got an error message when trying to do that step
Step 7: verify dongle.restore.go I don't recall being in the directions when I was going the test, but I rebooted
Post Reboot: None of my previous z-wave devices were listed. I also checked dmesg via ssh and noticed the following items:[ 4.328000] Unsupported Device! [ 4.328000] Vendor=658 ProdID=200 [ 4.328000] Manufacturer= Product=
I saw that item a couple times which almost seems like Vera is blocking the UZB1 or at least complaining about it.
I ended up switching the Z-Wave back to the embedded controller, and restoring configuration from backup.
Any suggestions what I did wrong??
Script to disable all the mios/vera proprietary program and broadcast its zwave and zigbee radio serial ports in your network so it can be picked up by another controller... For example z-way.rafale77/Nuke-Vera rafale77/Nuke-Vera
Script to neuter the vera and broadcast its zwave and zigbee serial ports - rafale77/Nuke-Vera
I have a VeraPlus and would like to purchase a second z-wave controler. Is it possible to build a second z-wave network next to an existing z-wave network? Slowly I want to move devices form Vera to the new z-wave network. Is it more advisable to install the second z-wave controller as a slave controler?
My VeraPlus is in Christmas Light mode after attempting the latest update. So as I attempt to get things back with Vera Support I need to start looking into an alternate solution. I've installed OpenLuup in an Ubuntu VM I have running on an old PC. Right now I can still ssh into the VeraPlus.Will I be able to migrate the Z-Wave NW to a UZB dongle in this state? Or can I use @rafale77 's Nuke Vera Script to just use the VeraPlus as my zwave antenna? With either of these options have I lost the ability to migrate my configuration without Vera support restoring my VeraPlus first?
So, after more than 10 years and 4 Vera's, my vera forum account is suspended.
I think the monkey of the company, called melih is not man enough te have any critic. In the topic about the "great to come app" I responded to the topic (not off topic!) that a subscription can also be seen (financially) as buying a box, run it for 2 years and come to market with a better one (in better wordings).
That means that if you as a company hast a box of 50USD cost, sell it at 250USD would give him a monthly subscription of 200/24=8.33 USD... Users that would last 5 years would pay a monthly of 3.33USD.
But... the monkey on the rock has bolded his fists and said: ban him! He just can't take any criticism.
To be honest! I am not even sad! It was a very rough journey with Vera. The journey stopped when Ezlo and the monkey took over the name.
After that it quickly went downhill...
Goodluck all! (except the monkey).
Edit:Aug 16, 2020 MiOS Application - early Beta enrollment MiOS Application - early Beta enrollment
are you serious? would I become a cost? that is, customers who periodically replace the control unit and are happy with the service praise the Vera system and therefore also free advertising but, the new “company” decides to charge less to bind the customer with a subscription … But the genius...
The point you are missing is:
If the hardware would “cost” around 50USD, and you sold it for 250USD, an average life span of 5 years would mean 200/(5×12)=3.33USD per month. Thats how Vera used to do it.
Then release a faster/better “beast” every 2 years gives you 200/(2×12)=8.33USD per month. That’s how vera used to do it.
And that’s how I did it for around 8 years and 4 veras with them untill you bought the company.
This is just a matter of how you look at it and explain it.
I dont frikking care how you do the maths, I only liked a working product and a close user community (like it was).
Now all I see on this forum is a lot of “we are going to” and “we are working on”, “keep them coming guys”. But the reason for ever considering to buy a new vera every 2 years is gone…
No clear migration path, no clear compatibility path and no clear integration path…
That’s why I started to look for alternatives…
I have 50ish devices in connected to Home Assistant via. VeraPlus hub. I recently picked up an Aeotec Gen5 USB stick which I have successfully integrated into Home Assistant using OpenZWave.
Can I add the Aeotec Stick as a Secondary Controller to Vera, turn off vera (making Aeotec primary) and then run natively that way on HA?
I know the entity ID's in HA will change, but a lot easier to fix vs. unpair/repair all the devices from Vera.
Will this work?
To answer @sdtoddl who emailed me asking how to manually force neighbor node setting on the vera, I don’t think it is possible with the UI.
The solution is actually to trigger a neighbor node update on the given device and all the devices located between that said device and the controller.
The vera deals with this by running multiple repeated (nightly) network heals which is not very efficient and causes it to frequently have unexpected network breakdowns and surprises good (some devices with problems get fixed) and bad (some devices which used to work stop working). It also takes down the entire network when it does it. One can however manually trigger neighbor node updates on the vera through the device’s advanced menu.
To go to the next level, this data, the network node list is actually not recorded in the host controller firmware but in the non volatile memory of the zwave chip both on the controller and on the devices. It is therefore independent of the platform you use and can’t be set by the host. One can only ask the zwave chip on the device to run a survey and then inform the controller’s zwave chip of what it has found.
How to overcome the vera's storage limitations:rafale77/vera-extroot rafale77/vera-extroot
Execrably poor partitioning and usage of the onboard flash causing hardware reliability failure through data corruption and storage availability for plugin limitations: Add your own external USB SATA drive and extroot.
Scripts to extroot the vera edge-plus-secure. Contribute to rafale77/vera-extroot development by creating an account on GitHub.
Amazing. One of my (un)trusty old VeraLite machines running UI5 has just achieved an uptime of one month... the maximum that that firmware can ever achieve!
The graph below shows the uptime for the last two years, and although sometimes being up for a few weeks, it's never achieved a whole month!
I put this down to reduced ZWave traffic and zero plugins, although the configuration hasn't changed for a long while.
Using Home Assistant as the heart of my domotica (this was vera for 10 years up untill just a few months ago!) I currently have vera integrated as zwave hub only. Having almost all logic moved to hass leaving vera solely as zwave hub.
As I am searching for a solution for a more stable zwave network and all options are open (but messing with zwave network ids etc is still too hard for me) I am also looking to another option.
Multiple veras in hass. This way I can have isolated vera edges with smaller zwave networks and still do all logic in hass.
One thing... currently only 1 vera is allowed in hass.
It is a feature request already and I would like to ask to help me vote this request up:Jul 26, 2018 Multiple Vera Hubs Multiple Vera Hubs
I have multiple locations with Vera hubs at each for Z-Wave/Zigbee controls. However, you cannot add two or more Vera hubs to your configuration.yaml. Please add this as an option.
The vera can run both its zwave and zigbee network off of external USB radios instead of the onboard ones. This could have advantages in terms of portability, facilitating migrations or recoveries in case the vera craps out. It could also enable testing of newer radio firmwares.
Zwave is pretty straightforward as its serial API is standardized and the protocol is the same regardless of what brand of USB stick you use. I have run my vera off of USB sticks of various brands for years going from Aeotec, Zooz, Homeseer and even the most generic silabs/Sigma Design. You just insert the stick in the usb port, find out which serial port it created (under the /dev/ folder) and use that port (ie. /dev/ttyACM0, /dev/ttyUSB0 etc) in the zwave advance menu.
Zigbee is a little trickier as the protocol and chipset was not quite as standardized as zwave. The vera only works with ember Znet protocol and you will need to find a usb stick with an EM35x chip in it which already has an ember zigbee firmware loaded. One example of such a stick is the Go Control HUSBZB1 (dual zwave-zigbee stick). The port is not readily accessible on the vera UI but can be accessed either through editing of the /etc/cmh/user-data.json or through ALTUI by going into the hidden zigbee radio device through the "table devices " menu and changing modifying the port variable there.
You can also upgrade the onboard zwave and zigbee radio firmwares from the command line SSH...
vera - motion sensor stays tripped
I know how to uptrip but it is so annoying that lots of motion sensor keep tripped...
("urn:micasaverde-com:serviceId:SecuritySensor1", "Tripped", 0, 986)
Any tips or suggestion to avoid this sensor behaviour?
HSD99 last edited by
I assume you mean that Vera is not seeing the "untrip" event from the motion detector, a sadly common problem in recent firmwares. I 've solved it by adding the "AutoUntrip' variable to the device with a \ time-out delay to untrip the sensor. 60 seconds is a good value, but you can choose any number you want.
Can you give me more information on how to achieve that? Very interesting!
I practically completely eliminated this problem after ditching the vera.
It is as @HSD99 says due to misses of the untrip frame coming from the device and this frame often lost because the vera is having fun with itself... either in a got CAN, in a tardy wait mode, reloading luup or that portion of the network doing a wakeup nnu or a wakeup poll. There are all kinds of mechanisms as you can see for it to fail and the larger the network is with battery operated devices and secure class devices, the more likely this happens. Hopefully this is not a ghost trip which then means that there is no untrip frame coming behind it... this is a vera data corruption problem either in the user-data.json file or in the vera RAM.
The auto untrip is one way to resolve it but it is really a workaround, patch... not one that I would recommend. The fundamental solution is to get the untrip frame from the sensor and to do this we need to reduce chattiness of the zwave network.
The autountrip is below:
local deviceID = **the device you want to set** local time = **the time after which it untrips in seconds** luup.variable_set("urn:micasaverde-com:serviceId:SecuritySensor1", "AutoUntrip", time, deviceID)
Yes, but since zway is not natively integratable in home assistant I am stuck with Vera... for zwave...
Short of going with z-way, you can go to the native openzwave on Home Assistant... Vera offers no advantage over OZW anymore. It has actually fallen quite a bit behind it.
And you don't even need new hardware. You can just "nuke" the vera and use it as your zwave and zigbee radio for home assistant using the home assistant components for them. In that case, you won't even have to rebuild your network. OZW will just take over the vera zwave network and make it this much better. You will have to rebuild Zigbee though.
I do have a vera edge, only zwave, no zigbee. And I would not know where to start before even "daring"...
Makes no difference that it is an edge
The zigbee part just won't work
There is zero risk in doing this. Just run a vera backup. Worst case, you just restore. I just provided a script to make it easier.
It's actually much less risky than starting a network from scratch since you don't risk not being able to include or configure devices and you can go back easily. Starting from scratch is a "no return" path.
Sorry, but for me this is rocket science. I do not fully understand it as a whole and therefor dont know where to start. For example, killing the processes except the ones listed... will that survive a reboot? The thing is, if I do something wrong I don't have enough skills to revert or go forward...
A bit surprising given the fact that you run home assistant which is much more involved and complicated to setup and maintain than this:
To answer your questions:
Does it survive reboot? yes... the last step of the process is to reboot.
How to undo? Factory reset and restore backup.
All it does is prevent all the mios programs from loading at boot up and adds ser2net which sends the serial signal over your network. You then get socat to pick up that port and reflect it on Home assistant as if it was a local serial port. You can thereafter use the openzwave component (zwave2mqtt, QTopenzwave, zwave, they are all based on openzwave) of your choice on Home Assistant. No different from your vera component today, just much better integrated.
Ok, I have a spare vera... I am willing to try...
So what should I do step by step?
Step 1. This vera has no zwave devices currenlty
Step 2. Nuke vera with you script?
Step 3. Follow home assistant link above?
Step 4. Have a zwave device connected to vera and it should show in hass?
What I dont get yet, I run hass core. A vm on esxi. I cant ssh to "linux" other than the basic linux on which hass runs but it seems I cannot really "do" something there. So how could I do:
On Home Assistant - first ssh connection
install socat with “apt-get install socat” or whatever package manager your distribution has...
And what I do not understand fully is how vera could then be an OZW1.6 one?
The Home Assistant link is a step by step. That one does not fully survive reboot.
My script run on the vera only does the vera part. It just goes a little further by making it resistant to reboot and disabling really everything from vera/mios. You still need to access your home assistant machine and install/setup socat per the home assistant forum. Maybe I should add that in my github read me to make it easier.
You have two possible scenario here.
- If you have no device on the vera, you will need to use Home Assistant to include them... yeah no more vera inclusions.
- If it is your vera you are migrating, all your devices will show in home assistant under the zwave menu if you use the zwave component, or the other zwave interfaces (zwave2mqtt or QTopenzwave) if you use one of the other 2 components. You can continue expanding your network from there. Again, no more vera software. You are just reusing the hardware.
Do your research and decide which one to choose. Home assistant offers many components. The newest one is QTopenzwave. The native zwave is the oldest but most integrated one.
I too run “Hass Core”. It’s really a misleading name because it is the most complete and flexible version of Hass. Not just the core. The others are just customized, limited versions to run as a container or a self contained OS. If you run Hass Core, you probably have a full linux OS in the VM.
As to your last question. Maybe this is where you are the most confused. What you call vera is primarily a piece of software. I described it in my post about the vera different layers of connections but I admit, does not really address your confusion. It is really just a program running on OpenWRT which is a Linux distro. No different than say Home Assistant running on ubuntu, z-way running on windows or any other OS.
The UI you see is an open source webserver program called lighttpd reading from the /www folder.
The Luup engine is a binary compiled from C code called LuaUPnP located in /usr/bin and it relies on a number of libraries and connects to the zwave radio through a program called zwaveserial which I talk about in another thread here. The zwave radio is connected to the mini computer through a serial port which is what the zwaveserial program calls. What you see on windows as a COM port is shown under linux as in the /dev/ folder as a TTYXXXX port.
Now the hardware. The various Vera Edge/VeraPlus/Vera Secure are devices sourced from SerComm, ａ company in Taiwan notorious for being an OEM for other companies producing mostly wireless routers and IPcams. The vera is hardware wise a wireless router with additional radios with its original operating system on top of which mios slapped a couple of programs. You prevent the mios programs from starting, and you end up with basically a device like a rPI with a zwave radio running OpenWRT linux. From here on you can think of it as just that. A tiny computer with your zwave radio onboard and forget that it is a vera.
Now zwave radio chips have the particularity to have onboard memory and this memory stores all the data from your network. It is made of the HomeID, which identifies your network, the list of nodes, your device list, and their routing. It does not contain the command class information for these nodes.
So... If you kill the vera programs, you end up with a tiny linux computer with a zwave radio and all your zwave network info in it. You can just send this serial port through ethernet to your home assistant VM and have home assistant manage it. This the purpose of folks in the link above.
Rafale, very clear and very understandable. But again option and options... as I am not an expert in this I will probably hit the wall after the first corner.