We have a number of Particle Electrons in the field. Recently we have had some stop communicating with our servers. At first we thought it might be this issue UDP Packets being lost between Network and Server. However in at least 2 devices (and maybe a 3rd), it appears that the keys are no longer valid. The LED blinks Cyan with the occasional red flash.
These devices are at remote locations and are expected to not need the maintenance that is required to correct the server keys issue. In fact we have incurred significant expense to even find out that this is the problem. This doesn’t appear to be repeatable.
Therefore we have 2 main questions:
What can cause the keys to become invalid? The particles were working fine for a number of months before randomly dropping off. Is there a way to correct the issue remotely?
The second question is if we change the system mode to MANUAL, will the key check and sync with Particle processes occur? We are thinking about doing this as we having problems maintaining communications in a static location and we think that having full control over the modem will help.
Can you state what system firmware the respective devices are running?
IIRC, the keys do get rotated frequently but the device does store two sets (current & next) for which the validity period should overlap long enough for next to become current and a new next will be provided.
One way to get keys out of sync is by not connecting often or long enough for the key update to take place and hence the last stored next to become deprecated without pulling in the new next.
But I have no info about how frequently the rotation happens nor how long the minimum period would be to get the key update consumed. Maybe @ParticleD could chime in on that.
To be sure, are you having problems with either the server address or server public key (fixed with particle keys server) or the device private key pair (fixed with particle keys doctor)? Do you know which one? Or both?
We are running 0.7.0 RC-6
particle keys server fixed the problem so I guess it is the server public key.
If the keys are rotated, then this is probably the issue as it happens when the devices go offline for a period of time. We try to reconnect as soon as possible, but this doesn’t always happen.