[Core] Cloud Connect Issue -- Recent changes to build server?

Hi all.

I’m compiling via cli particle compile and have recently found that my code is unable to connect to the particle cloud. I’ve reverted my own code back to a known good commit and the same issue continues.

The core (running in semi-automatic) connects to wifi, is able to talk to my own servers via UDP, but calling Spark.connect() fails and I get a red flashing LED (occasionally cyan flashing when the code tries to reconnect to the cloud).

The issue started June 30 around 4pm EST, just wondering if there were changes to the build server around that time? I have a binary built on Tuesday morning that seems to work fine when I flash it, but when I try to recompile from the same source I get the cloud/red-flash error.

Starting to feel like I’m on crazy pills, can anyone verify/deny changes applied to the particle servers on Tuesday afternoon? Starting to run out of ideas about what could be causing the issue.

1 Like

load tinker firmware if you are not certain and want to use a stable user code.

Building and :cloud: connection are separate things though so it shouldn’t be affecting :cloud: connection.

Yup, tinker works fine. So does a binary I have saved from Tuesday morning. Both connect to the particle cloud.

Since reverting my code back to an older git commit (that I know should work) compiles to have the same issue, I’m wondering if some source has changed on the build server. I don’t see any recent changes to spark/firmware or core-communication-lib on github. But I thought I would ask anyway.

I think @Dave made a change to fix another problem but I am not sure of the timing. Maybe he could weigh in here.

Could you be running out of RAM based on a built-in firmware change? I have seen firmware that is close to the RAM limit for one set of base firmware and over the limit for the next.

@bko Is there a way to get the output of arm-none-eabi-size via cli? Or get the object files so I can check it on my local machine?

I’ll check this out here, I think this is related to the photon needing a system update, I’m hoping to roll out the photon system update functionality soon.


1 Like


Just built everything locally, the issue is not there any more.
Building via build.particle.io DID produce the same error as well.

I don’t believe it was a size issue, flash / RAM are at 85K / 15K

I’m surprised no one else has seen anything similar, but maybe it’s related to running in semi-automatic mode.

Hey All,

Sorry about this issue, I’ve been looking into this, and it looks like the production build farm was misconfigured, and was pointing to a staging version of firmware instead of the production version. I am really sorry about this, if you’ve flashed a program to your core on the build IDE in the last day or two, your core may have been blinking cyan, and a reflash of tinker / factory reset should fix it.

I’m really bummed this wrong config change went out, and I’m looking into what I can do to prevent that from happening again. Sorry!



I haad the same issue and it happened after 5pm EST. I flahsed the device and everything went back to normal and I was able to load the example code. But then when I try to load my code, I am getting the same problem as before.

I am talking about this problem on this thread:

Thanks for looking into the issue, Dave.

Glad the issue is fixed. Had me losing my mind for a little while.
At least I got another anecdote to add to my rant about cloud-based IDEs :stuck_out_tongue_winking_eye:
and a good reason to make the effort to move to building locally.


Hi @jakeypoo,

Sorry about any distress there, I’m glad there are always lots of ways to build firmware, but I’ll never stop trying to make it easy. :slight_smile:


Was that just for the Core, or Photons also?

It should have been for cores, and just from last night around midnight to earlier today as far as I know.

1 Like