write() takes 5 seconds, because it gets stuck at http.getString()

[Cadence-GitHub]

I use the latest ESP8266-Arduino library (2.6.3). And the latest ESP8266_Influx_DB library (2.0.0).

When I execute the influx.write() call, it takes 5 seconds to complete. It gets stuck on this:
https://github.com/tobiasschuerg/ESP8266_Influx_DB/blob/5cf6cc93f6df73f18cebd6e06df79cc741002fe3/InfluxDb.cpp#L166

I use influxdb version 1.7.9.

[Cadence-GitHub]

Here is how the output looks like when I execute the code on my ESP8266 device:

17:08:05.407 ->  --> writing to mysensors:
17:08:05.407 -> sound,device=5,sensor=2 value=1015
17:08:05.407 ->  <-- Response: 204 ""
17:08:10.623 -> 1021
17:08:10.826 -> 1021
17:08:11.028 -> 1021

That "1021" is a sensor read that normally happens every 200 milliseconds. You can see that it got stuck for 5 seconds after sending data to influxdb.

[Cadence-GitHub]

Why do we even need this getString() here? Response code is what is really important here. Can we have a way to disable this getString() call when not needed?

[Cadence-GitHub]

Perhaps this is how it should be fixed:

if (httpResponseCode != 204) {
   String response = http.getString();
   Serial.println(" \"" + response + "\"");
}

[tobiasschuerg]

Feel free to create PR

[ivanpiffer]

I use the latest ESP8266-Arduino library (2.6.3). And the latest ESP8266_Influx_DB library (2.0.0).

When I execute the influx.write() call, it takes 5 seconds to complete. It gets stuck on this:
https://github.com/tobiasschuerg/ESP8266_Influx_DB/blob/5cf6cc93f6df73f18cebd6e06df79cc741002fe3/InfluxDb.cpp#L166

I use influxdb version 1.7.9.

I'm exactly in the same situation. 5 seconds lost :(.
Have you found a solution as a patch of ESP8266HTTPClient ?
Or any possible solition in 2020 ?

13:38:19.052 -> --> writing to amperometri:
13:38:19.086 -> mA,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=118.50
13:38:19.154 -> mW,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=483.00
13:38:19.222 -> V,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=4.08
13:38:19.290 -> <-- Response: 204 ""
13:38:24.553 -> Current: 118.40 mA
13:38:24.553 -> Power: 482.58 mW
13:38:24.587 -> Voltage: 4.08 V

[ivanpiffer]

I use the latest ESP8266-Arduino library (2.6.3). And the latest ESP8266_Influx_DB library (2.0.0).
When I execute the influx.write() call, it takes 5 seconds to complete. It gets stuck on this:
https://github.com/tobiasschuerg/ESP8266_Influx_DB/blob/5cf6cc93f6df73f18cebd6e06df79cc741002fe3/InfluxDb.cpp#L166

I use influxdb version 1.7.9.

I'm exactly in the same situation. 5 seconds lost :(.
Have you found a solution as a patch of ESP8266HTTPClient ?
Or any possible solition in 2020 ?

13:38:19.052 -> --> writing to amperometri:
13:38:19.086 -> mA,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=118.50
13:38:19.154 -> mW,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=483.00
13:38:19.222 -> V,device=amperometro,sensor=one,mode=pwm loopCount=863.00,value=4.08
13:38:19.290 -> <-- Response: 204 ""
13:38:24.553 -> Current: 118.40 mA
13:38:24.553 -> Power: 482.58 mW
13:38:24.587 -> Voltage: 4.08 V

### SOLUTION

I've overwritten libraries/ESP8266HTTPClient/src/ESP8266HTTPClient.cpp with the version of the file you can find in caf891d316 tree here (not still accepted in the master ?) https://raw.githubusercontent.com/esp8266/Arduino/caf891d316a689bc5564a88a7796d7a8496345f5/libraries/ESP8266HTTPClient/src/ESP8266HTTPClient.cpp

this is my output :) right now . ~ 1 seconds as my delay code delay(1000);
14:45:43.067 -> --> writing to amperometri:
14:45:43.101 -> mA,device=amperometro,sensor=one,mode=pwm loopCount=1113.00,value=118.90
14:45:43.169 -> mW,device=amperometro,sensor=one,mode=pwm loopCount=1113.00,value=484.61
14:45:43.237 -> V,device=amperometro,sensor=one,mode=pwm loopCount=1113.00,value=4.08
14:45:43.339 -> <-- Response: 204 ""
14:45:44.253 -> Current: 117.90 mA
14:45:44.253 -> Power: 480.53 mW
14:45:44.287 -> Voltage: 4.08 V
14:45:44.323 ->
14:45:44.323 -> --> writing to amperometri:
14:45:44.323 -> mA,device=amperometro,sensor=one,mode=pwm loopCount=1114.00,value=117.90
14:45:44.390 -> mW,device=amperometro,sensor=one,mode=pwm loopCount=1114.00,value=480.53
14:45:44.457 -> V,device=amperometro,sensor=one,mode=pwm loopCount=1114.00,value=4.08
14:45:44.559 -> <-- Response: 204 ""
14:45:45.476 -> Current: 118.70 mA
14:45:45.476 -> Power: 483.80 mW
14:45:45.510 -> Voltage: 4.08 V
14:45:45.544 ->
14:45:45.544 -> --> writing to amperometri:
14:45:45.578 -> mA,device=amperometro,sensor=one,mode=pwm loopCount=1115.00,value=118.70
14:45:45.645 -> mW,device=amperometro,sensor=one,mode=pwm loopCount=1115.00,value=483.80
14:45:45.712 -> V,device=amperometro,sensor=one,mode=pwm loopCount=1115.00,value=4.08
14:45:45.812 -> <-- Response: 204 ""
14:45:46.725 -> Current: 118.50 mA
14:45:46.725 -> Power: 482.98 mW
14:45:46.759 -> Voltage: 4.08 V

[thearduinoguy]

Will this solution mess up anything that uses the library?

[evanthomas]

I also hit this, although in my case it hangs indefinitely. Checking for 204 before calling getString() fixed it for me.

[ivanpiffer]

I also hit this, although in my case it hangs indefinitely. Checking for 204 before calling getString() fixed it for me.

Hi ! Can you show us your implementation? Thanks

[evanthomas]

@ivanpiffer I just changed the code in the function boolean Influxdb::write(String data) in file InfluxDb.cpp near line 166. I'm using platformIO so it recompiled it for me. Not whether the arduino IDE will do it automatically for you or not.

  String response;
  if (httpResponseCode == 204 ) {
    response = "";
  } else {
    response = " \"" + http.getString() + "\"";
  }
  Serial.println(response);

[tobiasschuerg]

This issue might be fixed with #26 and is contained in 2.0.1, please check.