HC3 Fault

[michal85pl 15]

Hello,

 

I have had a problem with my HC3 since some time. Every few days the device starts to work slower and slower. Simple scenes start to take longer and longer, a few then several seconds. When I go to diagnostics, the CPU load shows 100% first on the first core, then on every other core, and so it loops. After rebooting, the device loops and after a few reboots it goes into the auto repair mode or restore the previous configuration from backup.

 

Please login or register to see this image.

/monthly_2021_01/760446591_Zrzutekranu2021-01-09214242.jpg.1c396c32f3a9cb83f4d745b9391fbe8c.jpg" />

 

After manually restoring a proper backup, it works for a few days and the situation repeats itself. Sometimes I still get a notification about a hardware fault of the NICE module that I do not use. Does anyone have a similar problem and know how to solve it? Do i have to do a service request?

[SmartHomeEddy 878]

Sounds like you have something running, that eats up your system. Maybe shut down some scripts and QuickApps and see what is doing that to your HC3. 

[FBerges 52]

I had the same behavior.
For me, the memory usage was 96% and remained steadfast on the percentage even after a restart.
The CPU utilization was partially 100% for all CPUs.
As well as the slower and slower sequences of the scenes.
Some scenes weren't even started or ended, but I didn't notice it immediately because there was no error message.
I only noticed this after my blinds were no longer lowered.
I have submitted a support request.
After about 1 day I got the message Enable support access for 7 days.
After the Fibaro staff cleaned my memory, the memory usage was below 40% and everything was OK again.
After a while it started all over again.
Again support and everything was OK.
I cannot say why this is because the support did not provide any information.
What was also remarkable is that a Bachup in the cloud is only true twice possible.
After the adjustment, about 10 pieces were possible.
Whether it is due to poor memory management or the history, who knows.
The occupancy is currently 63% and is growing again.
I now use LUA to monitor the sequence of scenes and memory usage.
If something is not OK, I create a push mail.

Greetings Frank

[10der 831]

11 minutes ago, FBerges said:

I had the same behavior.

can you send to me response of 

http://192.168.1.36/api/service/logsDump

where 192.168.1.36 is your HC3

 

1) in personal mail

2) if you DO NOT upgrade your HC3 on the lates beta (sorry, i do not want to deep to this sh1t version right now)

[Sjakie 138]

Guys,

Just for information.

After some high COU load now 220 devices (Zwave and HUE)

Full HA by HC3

HC3 sometimes slowing down if more screens are open?

Arming is still in HC2 will be transfered shortly

JGab pointed me out I had some "hanging"API calls.

Please login or register to see this image.

/monthly_2021_01/image.png.10b430b327435a3d1353071dc59396f2.png" />

//Sjakie

 

 

[tinman 1,028]

1 hour ago, Sjakie said:

HC3 sometimes slowing down if more screens are open?

 

sure, each screen is doing data refresh

[Guest torneset]

I had some performance issues with my hc3 for a while. About 60 devices and when a scene is triggered it can take several seconds before action is performed, and other times it reacts momentarily. This is most annoying on the scenes triggered by motion sensors. I can see in the activity log that there is quite a delay from scene is triggered until device is responding. Support did instruct me to download the logdump, but I have only once been able to do that once right after a reboot of hc3. When it has been running for some minutes and I try to download from http://192.168.104.51/api/service/logsDump it reports 504 Gateway Timeout,

 

@10der - Have you ever seen this before? Do you know of other logs or methods to be able to see what makes the hc3 slow down?

 

I have also been checking for network congestion with a z-wave sniffer and I can not find any devices flooding the network. It seems like a healthy network with the right routing. And there is no difference whether the device is close to hc3 or far away (three floors in my house). 

 

 

[10der 831]

10 hours ago, torneset said:

@10der - Have you ever seen this before? Do you know of other logs or methods to be able to see what makes the hc3 slow down?

sorry, i have rollbacked to 5.050 after two days madness

[Guest torneset]

5 hours ago, 10der said:

sorry, i have rollbacked to 5.050 after two days madness

I had the latency problem on the latest releases including 5.050. No CPU peaks, lots of free RAM and still... periodically takes several seconds from motion to light is turned on.

 

Hope there will be some logging tools available not only for Fibaro support....

[10der 831]

2 hours ago, torneset said:

Hope there will be some logging tools available not only for Fibaro support....

this tools is Zniffer

[michal85pl 15]

12 godzin temu, 10der napisał:

sorry, i have rollbacked to 5.050 after two days madness

After today's crash and automatic repair to version 5.050, it seems that the latest beta was the cause ... for several hours on the restored version it is ok. On beta I couldn't even dump log - "Gateway timeout".
Strange because I thought it worked fine for a while immediately after updating.

[10der 831]

if LogDump doesnot work @tinman recommend to reboot your HC3 and trying again.

with a probability of 1/2 it`s help.

[Guest torneset]

Using Zniffer I see that all devices added in the early HC3 Firmware versions was added in secure mode. Did an incluison of a Fibaro Dimmer 2 yesterday and this was added in unsecure mode. Seems like I have to exclude/include all devices that I don't want to have running secure mode. Anyone seen this before?

 

BTW - I have also rolled back to 5.050 and it is much more stable.

[10der 831]

3 hours ago, torneset said:

Anyone seen this before?

in HC2 i was have a too much issues with sec mode with some devices. so never after connect device via sec mode.

it's my IMHO. opinion diverges from the pattern (maybe)

 

3 hours ago, torneset said:

I have also rolled back to 5.050 and it is much more stable.

still wondering why FIBARO is not rollback this FW.

@A.Socha  @m.roszak

is too much negative feedbacks and no reaction from FIBAR GROUP.  

[Guest torneset]

2 hours ago, 10der said:

in HC2 i was have a too much issues with sec mode with some devices. so never after connect device via sec mode.

it's my IMHO. opinion diverges from the pattern (maybe)

 

It was my policy as well, so when including all my devices in HC3 I made sure to NOT include in secure mode. I must say I was very surprised when I used the zniffer and every device was using S0 encryption.... When I do inclusion now the devices is included in unsecure mode, so it seems like a bug from the earlier releases of HC3. Now I have around 50 devices to exclude and include unless anyone know a trick to "convert" from secure to unsecure without exclude/include.

 

Agree with you that this Beta should have been taken offline...

[m.roszak 536]

There is no way to disable S0 encryption in included node, protocol does not allow it. 

As for adding devices always in S0 even if this was "unticked" in interface -> first time I hear of something like this and we didn't change anything there since product launch. 

As for beta, all reported issues are being investigated, if user experience issues with it it is always possible to rollback to previous version with 3 mouse clicks
I am not sure why you all think that we ignore provided feedback, we are working hard to fix reported issues as fast as possible.

[akatar 207]

3 hours ago, m.roszak said:

I am not sure why you all think that we ignore provided feedback

Well, that is easy, because fibaro doesn't respond to topics with problems?

[10der 831]

3 hours ago, m.roszak said:

I am not sure why you all think that we ignore provided feedback, we are working hard to fix reported issues as fast as possible.

 

Dziekuje, Michal!

[michal85pl 15]

7 godzin temu, m.roszak napisał:

I am not sure why you all think that we ignore provided feedback, we are working hard to fix reported issues as fast as possible.

maybe because 99% of our reports on the forum remain without any response whether it's positive or negative it's reported by the user and the releser ...

[xAPPO 18]

I was having real concerns over the future of Fibaro with the lack of response.  Seriously though your hardware always is leading but the software lags. You never seem to complete a platform release and HC4 will likely become available with 700 ZWave before HC3 is even half way complete.  You need more software development resource.

[akatar 207]

i think there will be a hc3 pro with a 700 chip (on short notice)