I have used Device Sync and recently deleted the app and redeployed it becauce of an authentication error related to the edge server.Now a new error has occurred.
When a session and connection end, the error occurs, and then another connection and session follow this step, seemingly to correct the error.
It looks like this:
The error message in the logs is:
error in read messages loop: could not decode next message: failed to get reader: WebSocket closed: write timed out: context deadline exceeded
Sometimes I also find a related error in Android Studio :
Hi, these are normal and transient errors for Device Sync. I will file a ticket to not surface them in the logs page.
The error indicates that the client has disconnected from the websocket connection and the server is attempting to read a message while that happens. This is not an issue for device sync as the protocol is designed to be disconnected and resumed at any point without breaking the synchronization.
Thank you very much. So the disconnection after a Realm transaction is normal? I usually see a disconnection followed by a reconnection shortly afterward.
Another question I mentioned before is about the Edge Server. Do you know about that? I have created a new topic, but nobody has replied.
I redeployed my app, restarted Device Sync and Edge Server, but the bug reappeared, I kept running the Edge Server on my Ubuntu Server, and sometimes its state turns to disconnected, and I can’t log in it as normal. The error message says:
ERROR: An error occurred: failed to handle command request "saslstart": invalid login request, full error: {'ok': 0, 'errmsg': 'failed to handle command request "saslstart": invalid login request'}
After I input the edgectl restart command, I can use it again.
Some hints I can see:
1.App Service Edge Server Last Connection stays unchanged, despite the fact that I have connected to it just now.
2.Lots of failed requests in the metrics, which disappear after restarting.
Hi, can you link to the other question? Would like to make sure it is answered there. The time issue is a known issue and actually fixed in an upcoming patch
