Error connecting to AMPQ: {:auth_failure, 'Disconnected'}

Hi

Almost every sequence I run over last week or so fails - most often due to this error:

Error connecting to AMPQ: {:auth_failure, ‘Disconnected’}

Perhaps @Marc or @Ascend you have some suggestions or ideas? Its not the most fun thing to do having to manually restart sequences where they got interrupted :slight_smile:

thanks

martin

…and I am also getting “sync errors” frequently when I press sync now button…

I had this error frequently before switching to ethernet. The WiFi distance was not too far, but the signal had to travel through my massive terrace ground which also became worse when I moved the garden furniture in that way. Also sync errors have become less frequent now, but they appear like once a week. I think this might be caused through a timeout for http requests. Sometimes it takes only 3s, but it also goes up to 5s and more.

Sadly its not that easy to install an external antenna to the Raspberry Pi. Maybe I’ll try that when I have one for spare next time.

I know exactly what you mean, when I had those issues I set my smarthome system to automatically ACK the e-stop and trigger another sequence try. That was probably not the intention of FarmBot, but Its really annoying if this comes up every day.

I had this error frequently before switching to ethernet.

I see, well I can test ethernet later this year, but I suspect in my case - wifi connection was working well for me since ~october last year. I still have signal strenght showing at around middle strenght / 67DBM. The only thing that got inbetween are the plants that grew on the bed :slight_smile:

I set my smarthome system to automatically ACK the e-stop and trigger another sequence try.

I am thinking myself about a more smart or even custom error handling. Maybe this could be a feature request? Nevertheless - maybe even farmware could do the trick - the idea was to:

  • Extend reporting of the error (which particular sequence failed)
  • ability to override locking mechanism for selected sequences - so that FB would retry execution of failed sequences, instead of locking himself. I would let him lock after e.g. 2 or 3 sequence run attempts
  • extended email with error reported - I would also add a photo from RPI camera to be able to see where / how did it got stuck…

If you can suggest a smart way to implement this or even have some plans for this in future - that would be great!

@Ascend, do you think this error is also linked to wifi vs ethernet issue?

certificate expired seems like some infrastructure problem rather that connection problem, does it not?