Photon shows offline, but its really online and running fine

We having some weird behaviors from the photon, It shows off line on the “particle list” command and on the web IDE.

In actuality the photon is running just fine, it is also communicating to our web database just fine as well. Meaning we are receiving HTTP POST messages.

This state seems random, but also slightly coincided with the turning on of another photon.

Any idea’s why this is happening ?

This is not a critical bug, just an annoying one as we use the particle list online status to help display active and online core in our website. The current status of this core is “Active offline” even thought it should be “Active online”.

What happens if you run other user code like blinky? Does it still show up as offline?

Having the same problem over here. Who is actually in charge of cloud software?

Am currently working with @corey on this.

@steelydev @Brian

we are working hard on this issue as it came in in the IOS and android apps as well. Devices would appear online as it sat in front of the customer but on the app it showed offline.

very similar to the readouts you are getting. Our primary cloud engineer is and already has been looking into this.

Please stay patient as we try to debug this issue.

-Corey

I’m sure it works, but who knows it seems random when these cores report offline but are online.

Are you talking about a core or photon?

Any updates on this issue? I’m experiencing the identical symptoms with a Photon (0.4.5) where the device shows offline with the app, but is indeed online and operating just fine, periodically POSTing to a database.

have you tried logging out of the app, closing it out, and logging back in? Basically clearing the app’s cached memory.

-Corey

Thanks @corey, that seems to have taken care of it. (Only curious thing is that my Photon is now being indicated as a Core on the app, but that appears to be an issue for this thread).

Thanks again.

@ido any reason this has been happening? I’ve seen this issue more than a dozen times in the last couple of weeks?

I am still experiencing this issue, How does the particle cloud know a photon is online? I assume there is some ping going between the photon and the particle cloud.

If I call a function to a photon that is showing offline, and it works shouldn’t that update the particle cloud that the photon is online ?

My post here is related to this thread. It explains our experience with the same bug.

Just saw this issue again today on a Photon running 4.7.

Hitting the /devices endpoint on the API I get:

...other devices},
{
    "id": "380XXXXXXXXXXXXXXXXX738",
    "name": "office_neo_ph_013",
    "last_app": null,
    "last_ip_address": "70.81.163.236",
    "last_heard": "2015-12-21T23:06:26.382Z",
    "product_id": 6,
    "connected": false,
    "status": "normal"
  },
{ other devices...

But hitting /devices/380XXXXXXXXXXXXXXXXX738 I get

{
  "id": "380XXXXXXXXXXXXXXXXX738",
  "name": "office_neo_ph_013",
  "connected": true,
  "variables": {
    "version": "string",
    "system": "string",
    "starttime": "string",
    "localtime": "string",
    "schedulelist": "string",
    "lastsync": "string",
    "boxorder": "int32",
    "boxtotal": "int32",
    "rssi": "int32",
    "ssid": "string"
  },
  "functions": [
    "transmit",
    "setschedule",
    "setvariable"
  ],
  "cc3000_patch_version": "wl0: Nov  7 2014 16:03:45 version 5.90.230.12 FWID 01-4310973",
  "product_id": 6,
  "last_heard": "2015-12-22T16:46:07.260Z",
  "status": "normal"
}

And both function calls and variables reads are working fine. I assumes the /devices endpoint keeps track of device status via online/offline events (and missed one?), while the device/DEVICE_ID endpoint hit the device directly…
Thoughts?

1 Like

Mine is doing this. Any update on this bug and how to move past it?

Also seeing this happen for several of my Photons.

Oh, I totally just created another topic with the same symptom. Oops. It’s at “devices” API is ill-informed on connected status

This remains a significant problem for us, and appears to have gotten worse with 4.9.

1 Like

This is a good question for @Dave

I assume there is some sort of caching going on between the …/devices end-point and the /devices/<> end-point.

1 Like

Hi All,

Thanks for reporting this, I wasn’t aware this was an ongoing issue, I’ll look into this. If you could send me firmware or details on reproducing this issue (listing all devices vs. get device attributes, etc), that would help.

Thanks,
David

2 Likes

Dave -

It doesn’t appear to be user code related. We have about 20 Photons regularly connected in the field, and 10 or so more in various test devices. We’ve seen this happen regardless of the user code they’re running. I believe I’ve even seen it with Tinker.

I’ve been logging the event stream. Several of our devices publish a lot of online/offline events, probably due to local network connectivity issues. The connection status discrepancy seems to occur more frequently in situations where the devices have been losing their cloud connection. I’ll be able to confirm that with data in a few days.

1 Like