Archived Forum

Hi

About a week ago my local Pi based emoncms that had been working perfectly for months suddenly stopped posting inputs.

With a great deal of help from Paul (pb66) I discovered I had 8000 individual rf_fail inputs that was clearly impacting performance. As far as I understand the number of entries was due to the rf_fail not having a nodeID. I believe the NodeID requirement has been added in a recent update so if you haven't updated you will not see the problem.

I altered this section of the NanodeRF sketch to add a nodeID and this, as far as I can test, has solved the problem.

if ((millis()-last_rf)>30000)
  {
    last_rf = millis();                                              
    str.reset();                                                    
    str.print(basedir); str.print("/api/post.json?");
    str.print("apikey="); str.print(apikey);
    str.print("&node=");  str.print(MYNODE);  // Add this line where MYNODE is the node no of the NanodeRF this sketch is uploaded to 
    str.print("&json={rf_fail:1}\0");                                   // No RF received in 30 seconds so send failure 
    data_ready = 1;                                                     // Ok, data is ready
    rf_error=1;
  }

Regards

Ian