I have a development imp that is getting the imp restarted, reason: out of memory. It has happened 3 times since the v30 firmware update. The imp remains off for 100 to 105 minutes before reconnecting and posting the error to the IDE. This is happening about once every other day.
If it is helpful to know what the imp does, it collects data and sends to the server once per minute. During the minute, it deepsleeps every 5 seconds between readings and stores the reading in ‘nv’ of getroottable() by incrementing the value (nv.data). After sending the data, I set the nv.data = 0; to reset the stored values.
When gathering code to send you, I noticed I call imp.onidle(function() { imp.deepsleepfor(READING_SECONDS); }); twice. I removed one and have two imps that have been running well since then. I’ll let you know if I find this is not the problem. Thanks!
That shouldn’t have caused an issue; there can only ever be one onidle() handler, and a second call would just replace the first one. If you replicate any problems though we’d love to see the code.