erizoController and nuve crashing

We have a licode installation on an ubuntu server in AWS. If the server is unused for an hour, when we attempt to reconnect again, the erizoController.js immediately crashes. I’ve also seen nuve.js crash, but I’m not sure it happens every time. The logs for erizo_controller and nuve contain only two lines:

[INFO] ErizoController - message: erizoClient connected, clientId:
[WARN] CloudHandler - ErizoController 0 in does not respond. Deleting it.

Anyone else seeing this?

Any suggestions about what logging configuration or other settings I should change to get more information and diagnose the problem?

I was going to set up the node processes with pm2, but if they don’t crash until I attempt to connect, it seems like I’ll get a failed request when I restart (at least one of them).

I’m experiencing a similar behavior in two of my servers. The logged error in my app, when creating the user token, is “No Erizo Controller found” but there is no EC related error in the Licode logs. I’m using pm2 to monitor processes but it seems it does not recognize the crash and it shows the EC process running. After restarting the EC every thing is fine.

It was good if Licode uses individual log file for errors. Finding failures in a bunch of log files is exhausting.

This is consistent in my case.

Note that we also hosted on Amazon AWS Instance.