Connection to the server

mdemmerle

I have been working with repetier server for a week, and my installation includes 6 printers spread over 2 raspberry pi 3. I regularly have a connection problem with both servers: i have the following error message "Connection lost. Oops ! i lost my connection to the server. Trying to reconnect".

This error message appears at the same time on both servers. It can last a few seconds like thirty minutes. The problem does not come from the internet connection because it works very well during the server malfunction. I also thought it was a Raspberry alimentation problem, but they are still connected to the internet box during the problem. So I have no idea what the source of the problem is.

Repetier

The message appears when the websocket used for connection gets no response for some time. This can happen
- when server crashes
- server pc or your pc loose connection to your intranet (or pi switches to AP mode)
- a command takes longer then the internal timeout for websocket
- Browser send tab to sleep so it did not send ping quickly enough (browsers throttle inactive tabs)

What is strange is if both loose connection at same time. Never heard that so far. So question is:
- can you connect with ssh at that time?
- does reload in browser help to restore? Would mean server is ok just websocket was dropped and not reconnected. Also we normally try every few seconds in case if interruption.
- Does a print continue? Means server it self did not crash.
- Anything configured that requires a connection to same service? If that would take long it might block websocket until it times out or has success. MQTT/Webservices are meant here.

Since both loose connection same time I assume there is one common part that is causing it. Any idea what that could be in your case?

Using
https://www.repetier-server.com/knowledgebase/debugging-crashes-hangs-on-linux/<br/>with backtrace on a server that is currently hanging it would be possible to identify it most likely if all is running. Just be warned that print will stop and will possibly not recover from staying too long on same position. So that is best done when you have connection lost without prints running.