Is Fibaro prepairing to introduce new gateway soon?

[Sankotronic 2,370]

Hello all!

 

As many of you know I'm developing UHAS (Universal Home Automation System) for Fibaro Home Center 2. This is complex system that should make Fibaro users life easier and give them decent HA without needing to dig into the programming in LUA and also should support 27 different languages with automatic translation of VD's. I use that same system on my HC2 and till firmware 4.140 it was working great. System was responsive and even CPU usage was a bit elevated it was not making my box sweating.

 

After introduction of FW 4.170 everything fall apart. That system was first to introduced lags in the system and also CPU usage get more wild. It was first FW that FORCED me to make complete RECOVERY due to never ending "Starting services". Before 4.170 all problems could be solved by either rebooting or restoring latest backup.

 

Latest betas 4.503 and 4.504 continue to make HC2 less good but introduced new backup system. To make story short, new firmware upgrades are making HC2 less capable but in the same time new backup is introduced that will make it possible to restore it to another gateway.

 

This looks to me like Fibaro is trying to copycat Apple. Apple practice is to introduce new updates to older devices that usually make them slow and annoying forcing users to buy new devices. Is Fibaro trying to do the same?

 

My experience in short with my system (HC2 with 101 included z-wave devices:

• 4.080 - started to develop UHAS since first preliminary tests proved that this is doable
• 4.100 - Introduced icon problems so had to rewrite UHAS code that handle that part
• 4.120 - found problem in UHAS code that could make gateway stuck with API calls due to way how HC2 was handling non existing icons in VD's. Problem could be experienced only after recovery when icons are deleted and backup is restored
• 4.140 - More than 50 UHAS compatible VD's and almost 20 UHAS compatible scenes are running smoothly on this system with slightly elevated CPU and memory usage. System runs for more than a month without any problems.
• 4.170 - Same UHAS configuration as on 4.140 is making CPU usage berserk (crazy, more elevated) and lags are introduced to system. Lags are of different length, from 1 to 60 seconds. All z-wave activity is stacked in memory and can be seen happening on GUI but with delay. So it is not z-wave that has lags, but system is not capable of reading all z-wave activity so it is stacked in buffer and then read with lags.
• 4.180 - same as 4.170. This one should only solve remote connection problem introduced with 4.170
• 4.503b - more resources are eaten by the system leaving UHAS less space to work. CPU usage is so crazy and jumping so much that I don't even believe it any more. lags in the system get worse or at least was not solved. Also introduced some new debugging like API: Not found etc. which definitely introduced more problems to LUA execution speed. I have completely rebuild my z-wave network on this FW to avoid any trouble caused by older z-wave devices. I can tell you that Fibaro motion sensor 2.8 and D?W sensor version 2.5 are not properly included in this FW
• 4.504b - All the same. System can became completely unusable due to long lags in the system where z-wave is responsive but system is getting all z-wave activity with lag.

 

At the moment I have no idea what to do next. I have freshly builded HC2 with included 101 z-wave devices with some decent code but doesn't work! Fibaro what is your next move?

 

[Bodyart 231]

Hi @Sankotronic,

 

you took the words right out of my mouth. With each new FW the more resources are used and more lags are appearing

Fibaro is hiding their roadmap from users, so we do not know their game, but the current situation is very disappointing, because our HA is failing, freezing, and making the users helpless and unhappy.

[gucio1234 79]

I agree with @Sankotronic , @Bodyart and others who suffer with firmware more recent then 4.160.

 

Dear FIBARO,  I hope you read this forum. Please take the comments on this forum seriously.

 

Lots of people invested lots of time, effort and money to build up our automated homes. I believe we want to continue and rely on your products ...

but for us to rely on them, they have to be reliable.

 

FIBARO - wee need to hear your comment on what @Sankotronic described starting this thread.

 

[petergebruers 1,268]

1 hour ago, Sankotronic said:

I have freshly builded HC2 with included 101 z-wave devices with some decent code but doesn't work!

I've read your other post and as I understand it, you did not exclude and include devices. Correct?

 

I'm not saying that this is your cause, but I think it is an important thing to know.

 

Frankly, I have given up trying to help people with "delays".

 

It is nothing personal.

 

It is just VERY difficult. I have tried to help quite a few people, sometimes via PM. It takes an awful lot of effort, from both parties. Sometimes it worked.

 

I think one of the main problems is the lack of diagnostics and tools. You mention CPU... which process/subsystem/scene is eating your processing power? I have been asking for diagnostic tools, statistics since 2013 and it is all locked away. I even made a few suggestions on bugzilla. Remember bugzilla?

 

That being said, there are indications that 4.170 and later are "slower" (or at least parts of the HC) than previous versions. I use quotation marks because what seems slow might be the result of so many things (optimizations, locking, different architecture, extra checks but in case of iPhone 6 deliberate slowdown) and might in fact be necessary to avoid other issues. I never found a good explanation.

 

All I can say @Sankotronic is I am sorry to hear it is such an ordeal. We've had many good discussions and I've got a feeling you are close to giving up.

[speedy 64]

1 hour ago, Sankotronic said:

 

• 4.140 - More than 50 UHAS compatible VD's and almost 20 UHAS compatible scenes are running smoothly on this system with slightly elevated CPU and memory usage. System runs for more than a month without any problems.

 

The "Cannot query interpreter state" was introduced here after 4.140. (in V4.150 in november)

Before 4.150 all my customers systems have been as solid as can be expected and all functions working as expected.

4.150 and over is where the system really lost stability and has not been fixed since.

 

Going back to 4.140 solves the issue every time, but this is not a sustainable way of dealing with it.

 

Quote

• 4.170 - Same UHAS configuration as on 4.140 is making CPU usage berserk (crazy, more elevated) and lags are introduced to system. Lags are of different length, from 1 to 60 seconds. All z-wave activity is stacked in memory and can be seen happening on GUI but with delay. So it is not z-wave that has lags, but system is not capable of reading all z-wave activity so it is stacked in buffer and then read with lags.

 

 

Same experience as you with multiple systems, the lag, freeze, interpreter state, whatever you want to call it.

There was something really nasty introduced in 4.150+..

 

And its not z-wave delays due to user error, its something wrong with the systems that causes the delays!

Updating from 4.150 to 4.140 introduced massive delays in some systems and going back "poff" the delays in z-wave traffic where gone instantly.

 

 

Edited July 23, 2018 by speedy

[Bodyart 231]

38 minutes ago, gucio1234 said:

FIBARO - wee need to hear your comment on what @Sankotronic described starting this thread.

 

+++++

[petergebruers 1,268]

@speedy that is what i thought as well, but as an end-user I do not have enough authority to say that. I think trouble started when that "Cannot query interpreter state" appeared and when they fixed that the CPU problem got worse...

[Guest spazpeker]

Hi @Sankotronic

Like a few others i have been using node red to do all the clever stuff, if you have a play with node red i think you will find that all the hard work you have done would be easy to transfer to any controller, thus avoiding the Fibaro dead end road

i now use node red for alot of alexa, finally i have managed to control my screenline blinds (node red via rfxcom)

you just push and pull info out of the HC2 and connect to my nest

and i can use Alexa to turn on lua scenes

Its like being let out of a cage !!!!!!

 

 

Please login or register to see this attachment.

Please login or register to see this attachment.

 

[speedy 64]

22 minutes ago, petergebruers said:

@speedy that is what i thought as well, but as an end-user I do not have enough authority to say that.

Well, i think you have an experienced opinion as you are very experienced in Fibaro  

And also, sometimes, even with no authority we can make qualified conclusions based on experience.
 

Update to 4.150+ and experience delays, freezes, "Cannot query interpreter state" and go back to 4.140 and its gone, its an open and shut case to me  

 

Quote

I think trouble started when that "Cannot query interpreter state" appeared and when they fixed that the CPU problem got worse...

Exactly in line with my own experience, Fibaro have been stable enough for me with 20+ systems until that error was introduced.

I dont think its fixed at all, the message is gone but the freeze (that i think introduced the message) is still there, and i think many people have problems but are not "techy" enough to understand it properly or are not using the system in the way that introduces the problems to them.

 

Edited July 23, 2018 by speedy

[Highlands 0]

As a new user, who recently purchased a HC2 and is only now realising the huge amount of work ahead of me in learning, coding, designing the HA system...I am reading these threads describing serious issues, and thinking, "Is it wise to even begin this journey" when such prevalent system bugs and issues will knock you down regardless of your own skill level, as I am seeing above with the some of the products most experienced users.

 

This is going to take a HUGE amount of my time to get  up to speed, I was told "it just works" by the resellers pushing this product. - I am in the very early stages, so really need some assurance that this product is the right one to begin investing my time in.

 

 

[petergebruers 1,268]

5 minutes ago, Highlands said:

I was told "it just works" by the resellers pushing this product.

 

Don't panic.

 

My Z-Wave rules of thumb: less than 5 modules, sometimes issues because not enough devices to cover all areas. Between 5 and 20...almost never issues. Between 20 and 50... sometimes trouble. Between 50 and 100 devices real "weirdness" starts and communication gets "fragile". Above 100 make sure you have the tools to diagnose network issues (ie sniffer, you can make one yourself) or get in touch with a good certified installer to support you (I am and end user, not an installer). Know about CPU and memory usage. Follow this forum. I have participated in many topics regarding "delays" and it is never easy...

[Sankotronic 2,370]

17 minutes ago, petergebruers said:

I've read your other post and as I understand it, you did not exclude and include devices. Correct?

 

Hi @petergebruers ,

 

first of all I want to thank you for your willingness to help and I have great respect for your knowledge.

 

To answer to your question:

 

I have completely clean my HC2 with doing recovery to 4.503 beta. I ended up with completely empty system as it came from factory. After that I have first reset all my mains powered devices and start their inclusion with the plan, from nearest to HC2 toward the ones further away. After I included all mains powered devices (smart plugs, relays, dimmers2, roller shutters, sirens etc) I started with adding battery powered devices but I left door/window sensors to be last included since older versions have problems with resetting and inclusion to the network. Last included where D/W sensors version 2.5 and three Fibaro eyes version 2.8 which proved to be difficult to include. More details about that I can write later.

 

Anyway after all 101 devices where included I did some mesh reconfiguration mostly for devices that are far from HC2 and don't have direct communication with HC2. All reconfiguration was successful and I always get positive response. After that I took some time to monitor my z-wave network and there was no delays in either sending command or receiving reports. Everything was almost perfect so I made local backup to be able to go back to clean system with only z-wave network.

 

Why almost perfect? Well, I was getting some Application rejected notifications for different devices, but mostly for one of the Fibaro smart plugs. Also I noticed that RGBW modules are struggling with reports and sometimes they act weird on the GUI by changing from on to off if program was running and then back on with white color setting and then back to off. Of course in reality RGBW module was just turned off and was never turned on with color and then switched back off.

 

I have also included 3 Fibaro eyes v2.8, but inclusion didn't went without problems. After first inclusion they where without battery sings like it is mains powered. Only after several attempts I manage to get battery sign but only on motion sensor slave and master while all others remained without it. Also after inclusion temperature and lux sensors didn't work immediately and both show 0. After several wake up the start reporting temperature and light measurements.

 

I have also several Fibaro D/W sensor version 2.5 which proved to be very difficult to reset and include even are positioned very near HC2. I was getting a lot of learning mode faults or where included as non configured device. One of them is now reported as dead, but if it is triggered by magnet then it get normal again. I still have to check but I suspect that it doesn't wake up at all even wake up interval is set to default 4000 sec. This D/W sensors is only 3 meters away from HC2.

 

As i was installing back my scenes and DV's I noticed after some time that lags are back. I do use extensively global variables in my code, but that approach did work very well on 4.140. I have 56 virtual devices with main loop running every 10 seconds and only checking if there is any changes in their setup. There is no API REST calls just opening few global variables doing checks and then waiting for next run. Of course I have 22 of VD's that are making HTTP request to HUE bridge every 10 seconds and then updating VD labels and sliders, so nothing out of the ordinary.

 

I have only one scene that is running all the time, Main scene FTBE, that also checks some global variables, do updates to timers and also runs some scenes or press buttons on VD's at repeating or scheduled time. I have also 15 scenes that are controlling lights and they run when necessary triggered by motion sensors and sending commands to devices or pressing buttons on VD's only if necessary so they are definitely not making a lot of traffic on the network. The rest of the scenes are run only at predefined time or when needed.

 

All above worked fine on 4.140. System was stable for weeks and there was no delays or lags. All trouble started with upgrading to 4.170 and upgrade was the only change made to that stable system. I was using another HC2 for my UHAS development and testing at that time.

 

54 minutes ago, petergebruers said:

All I can say @Sankotronic is I am sorry to hear it is such an ordeal. We've had many good discussions and I've got a feeling you are close to giving up.

 

I hope that we will have more good discussions here on forum. Also I'm not easily giving up and I will continue to 'fight', but it is really sad to see so many problems that users have and there is no way to find out what is wrong. I totally agree with you that it would be perfect if Fibaro can add some kind of diagnostic tools to their flagship product. What is now on diagnostic panel is ridiculous since it is also affected by the problems in the system. I really don't trust that CPU usage panel, maybe only memory usage is OK but is definitely not diagnostic tool, it is just nice info to have. Fibaro where is the rest?

 

 

[petergebruers 1,268]

@Sankotronic thanks, everything is very clear to me now. Sorry, no idea where to start troubleshooting. Except, maybe... try to persuade Fibaro to work with you to diagnose your system. I know they are capable of doing that, I once visited someone with the first version of Dimmer 2 (random flicker issue) and they spend *hours* in a remote session, working with the customer (for visual feedback...). BTW that D/W sensor is indeed difficult to include and I have one "dead" device but it is reporting status, just like yours... Not sure about the FGMS 2.8.

 

37 minutes ago, Sankotronic said:

Also I'm not easily giving up and I will continue to 'fight'

That makes two of us.

[speedy 64]

3 minutes ago, Sankotronic said:

I have completely clean my HC2 with doing recovery to 4.503 beta. I ended up with completely empty system as it came from factory.

After that I have first reset all my mains powered devices and start their inclusion with the plan, from nearest to HC2 toward the ones further away.

Wow, having to go to that extent to fix a system, even if it gets fixed is just terrible.

Think of installers with possibly 100+ installations at customers having to to that because of a firmware update, not counting all scenes, VD, variables, etc. etc. etc that need to be restored!

Point is, this should not have to be done at any point in time... Ever!

 

3 minutes ago, Sankotronic said:

All above worked fine on 4.140. System was stable for weeks and there was no delays or lags. All trouble started with upgrading to 4.170 and upgrade was the only change made to that stable system.

This is a: Maybe its not me, its YOU.

If your partner doesn't want to change its time to leave, how will you go forward with your development with UHAS knowing that this might happen again even if this gets fixed and you are then at the mercy of Fibaro with customers that have bought UHAS screaming at you...

I really want Fibaro to fix the problems and use the system (as i am an installer), but after almost 8 months! of not fixing this i hope and pray every day that something else comes along that have the same functions with the same ease of use.

I love the Fibaro hardware (dimmer V2).

 

I have stopped recommending the system, and have no faith in installing it as i know it has problems to big to overcome with code as its in firmware.

Before 4.140 there was some caveats that could be fixed with code or workarounds but the system was mostly stable and usable.

Fibaro have so many things good, but the fundamental stability of the system is the key to everything, why this is not understood and first and second priority with Fibaro i will for my life never understand!

 

Our reputation as installers lie in the systems stability, and Fibaros reputation lies in the hands of their customers.

 

It can only go 2 ways but i would like: Listen to people with problems, fix them with the highest priority and transparency, fast and you will be a company to trust with a fantastic future.

 

[chaicka 195]

Pardon me for pouring cold water... I have not been able to try 4.50x betas yet as it has been pulled. From what I see in posts, it is far better than those of us who embarked from 3.5x/3.6x to 4.0xx betas. The initial 4.x betas were a total nightmare...total unusable system can happen any time without any specific reason/rationale, corruption, endless exclusion and re-inclusion, etc.

[HomeSystem.sk 69]

2 hours ago, Sankotronic said:

 

Hi @petergebruers ,

...

 

Oh shut, so the complete system "restart" did not help? This actually makes it clear that this issues are not connected to Z-Wave as a technology but to Fibaro as a system. This is really bad news, I was really hoping we are out of the bad bad 5.180 story  

I really wish Fibaro would sit down together as a team, look at it and give a clear answer to the clients / installers. Now it is a bit tricky to even sell it if you are almost sure it will have lags... The question now is, if Fibaro knows about the issue (they know, but if they are opened enoguh to see the issues) and if they will be able to come out with update that fixes it without the need of "restarting" entirele systems. Just so you understand, we are at the point where if "restart" is needed to get the system back in track, we would have to do it for many many systems (maybe 30-40). That really is not a joke to me, as obviously customers would not pay for instalation again and Fibaro wont care...

[speedy 64]

17 minutes ago, HomeSystem.sk said:

Now it is a bit tricky to even sell it if you are almost sure it will have lags

Yes, I've stopped recommending and installing it since this happened in 4.150, as i will never be 100% that it will work any more.

 

17 minutes ago, HomeSystem.sk said:

Just so you understand, we are at the point where if "restart" is needed to get the system back in track, we would have to do it for many many systems (maybe 30-40).

That really is not a joke to me, as obviously customers would not pay for instalation again and Fibaro wont care...

Amen!

 

I dont know how many hours (40h+ for free!) I've spent with only 1 customer redesigning (working around stuff in LUA) in his system after 4.150 update and "Cannot query interpreter state" happened!

He has 135+ physical devices in 1 of his houses and its my biggest customer and i recommended Fibaro and he now controls everything including alarm.

 

So in reality, Fibaro is making me loose money.

I dont want to think about it, its money lost because of Fibaro and he now regrets ever installing it (and he will not buy more stuff or recommend it to others).

[jimicr 90]

I feel sorry for @Sankotronic, he has a job that takes him away from home for weeks on end.

When he's home he's losing family time because of this nonsense. And still always willing to help anyone.

 

I really hope Fibaro fix their systems, sometimes the lags are so long that when I get something from the kitchen, walk back to the couch sit down and THEN the light turns on in the kitchen.

 

Jim

[Joep 200]

Fibaro was busy selling their brand: 

Please login or register to see this link.

I hope this is not the end of the HC2, because I've just invested al lot of money in it. 

[AutoFrank 372]

I'm running 4.180, no CPU, memory or storage resource issues

• I have noticed delays in scene execution (up to 10 seconds) with no recognizable pattern with scenes impacted or date/time of delay
  • This also impacts my system where I 'sequence' two or more vd and scenes into an execution pattern. The lag impacts the expected 'wait delays' that I have coded in
• I have seen VD main loop crashes with json related "decode" errors that I have never seen before
  • Like others I have a VD detect/restart that takes care of restarting these
• I have noticed delays when I am trying to execute a scene manually or start a VD debug console
  • It can be several seconds or long enough that I question if I pressed the button at all
• I have noticed a couple of VD's that "stop" but have no error in the debug screen and sometimes the debug screen is blank
  • Much more difficult to detect and correct
• I have had multiple occurrences of what my family call "flash-overs". This is where all the lights will come on at once
  • This has also manifested itself where a scene will trigger a relay that is not referenced in it,
• Finally, the web UI has never been very good at providing real-time feedback but this is getting worse
  • I see this most often if I 'open' a VD and look at the labels. I frequently have to refresh the browser page to see the change
• My android app has  a serious difficulty in maintaining a local connection to the HC2.
  • Many time I need to switch off my phone wifi and connect through the fibaro cloud over cellular

I do not think it is z-wave related (even thought the newer generation of zwave devices are better at keeping the mesh intact )

I am convinced that it is system/HC2 OS related but have given up trying to understand the causes and just code or work around the issues