Connection Timeout

rajesh.palanisamy

Hi,

I am using REDUX dev tools in my angular web application, there is spiked data traffic on snapshot retrieval and the plugin crashes with the below logs and reconnect happens,

INFO |L.requests |PUMP POOLED THREAD 1 |Closed session <some sessionID> with internal cause code: 38/2 (Connection interrupted by client).

INFO |L.requests |PUMP POOLED THREAD 1 |Closed session <some sessionID> with internal cause code: 39 (Timeout waiting for client rebind).

Questions:

1. Is there a way to fix the connection issue since the reconnect pulls the redundant data.
2. Can I get the error code mapping - 39 , 38/2 .. etc for understanding purpose

Thanks
Rajesh

Dario Crivelli

Hi Rajesh,

About 2:

There is no detailed documentation of the error codes, so the last part of the log message is the available description of the code.
In fact, the "cause codes" of session closure are provided just in case they may be useful.

A session can be interrupted for various reasons; some of them may be problems while others may be just the client's will.
In your case:
"Connection interrupted by client" means that the session was streaming but the Server found the socket closed by the other peer;
"Timeout waiting for client rebind" means that the session was currently not streaming (the session migth have been either in long polling or still in the startup phase) and the client was supposed to perform a request but didn't.
In these cases the Server cannot even determine if this was the intention of the client or a client issue.

About 1:

Please clarify how you would like to cope with the issue.
Since you say that "the plugin crashes", it is difficult for us to figure out what happens.
Does the "crash" involve Lightstreamer client, and is this the cause for the connection interruption?

May you please expand on question 1, which perhaps can help us understanding the scenario?
In particular, what do you mean by "since the reconnect pulls the redundant data"?
Do you mean that, when restarting after the crash, the part of the snapshot already received before the crash would no longer be needed?