Having trouble seeing where I can control the behavior of the Imp during a WiFi outage. I am seeing this in the log:
[Exit Code] imp restarted, reason: wifi outage
I want to be able to execute code on the Imp during an outage, but the Imp keeps rebooting when the WiFi goes down. How can I deal with this?
Also, is there no function to determine if there is an active server connection? There is device.isconnected, why not agent.isconnected?
Thanks -AP
Sorry, I see the server.isconnected() now. First question still stands.
You should change your policy to RETURN_ON_ERROR and optionally install an onunexpecteddisconnect() handler.
I did do that, same result. I still get a reboot and that message in the log. What I am doing is turning off the router or removing the WAN connection from it. Either will trigger this behavior.
? That doesn’t seem right. Could you post your code, or PM it?
Hi was there a response to this issue as I am facing the same. I have implemented RETURN_ON_ERROR, but the device disconnects and then after a while I get the ‘Restarting due to wifi outage’ sequence which boots the device from start.
@rajkrishna this is fixed in release 30 which is going to RC any moment now. Wifi errors are dealt with seamlessly under the hood by rebooting the wifi chip and re-establishing the connection.
Where can I get release 30? Or it is a hardware release. Do I have to buy a new imp??
@ottelo no need for new imp your firmware will be automatically updated by the electric imp cloud service that one reason it so good 
BTW we are all awaiting release 3.0 sounds like it’s soon, that is a @Hugo SOON…
If people would like to try 30.1, then PM me your mac address and I’ll add you to the list…
I guess its time to get a DJ wash panel (maybe this one), and write some DMX code for Kaylee. I wonder if I can make a Batman beacon with just LEDs… hmmm…
Mi imp upgraded its firmware last week and today it was 10 minutes offline due to wifi outage.
A new version of 30 is going out to developers tomorrow with some fixes; our testing of these has gone really well so hopefully that should be the last of the wifi outage issues.
Just to add to this I have two imp that now need power cycle once a day after wifi outage. but not every outage it should be said. Is that all developers of just those seeing issues?
It depended to some extent on the code the VM was running, which is one reason why our normal testers didn’t see the issue. Certain types of code would exacerbate the issue. 30.15 will go out later today!
Hugo on boot I go to deep sleep and either wake every hour or on Pin1. Code hasn’t changed in over a year.