HC2 slow to respond to any command from anywhere

[h2009 0]

Hello 

I’m having a issue which seems to be getting worse. Basically the web interface and the iPhone app all respond quickly (no slow down) however any trigger of a device ie light switch turned on the system can take up to 3 minutes to turn it on if it responds at all. 

 

If if I reboot the HC it seems to work fine but then it after a day or so it slow down. 

 

All processes on the HC show within normal ranges, no heavy loads and all scences are disabled. 

[TRicky 84]

On which firmware are you working?

I'm experiencing familiar problems on FW 4.161 Beta.

[sir_horn 2]

13 hours ago, h2009 said:

Hallå 

Jag har ett problem som verkar bli värre. I grund och botten svarar webgränssnittet och iPhone-appen snabbt (ingen saktning), men några utlösare av enheten och ljusströmbrytaren slår på systemet. 

 

Om jag återstartar HC verkar det fungera bra. 

 

Alla processer på HC-showen inom normala områden, inga tunga belastningar och alla scener är inaktiverade. 

How many devices do you have? I had the same problem before, had about 180 devices on one HC2. At that time all virtual devices worked perfectly but z-wave funvtionality was extremely slow. I had a lot of contact with support who instructed me to a lot of different small changes to not put to much load on z-wave ship such as:

 

-setting wake up intervall to 21 000 seconds on battery devices

-disable polling on all non battery devices

-change main loop time on VDs (was about 1 minute and works fine for me with 30 minutes)

-I also changed motion sensors parameter 1 to 25, parameter 2 to 15, and parameter 6 to 750

-run a restart scene every night

 

Above did improve the overall performance but it was not enough and since I was going to add a lot more devices I decided to rebuild the system with another HC2 as a slave. Since the rebiuld it works as a charm and I have also added more RAM

 

[h2009 0]

Hi All, 

It was on the current up to date firmware also on the current Beta. 

 

I have 44 devices on the system so its not much to be creating this issue. 

Polling Dead devices is = off 

Polling time = 300s 

Mark dead nodes = on 

 

Ram = 56% free 

CPU load = 8% each core