Flaky test: `WebSocketServiceHttp2RequestCompleteTest.closeAfterTimeoutWhenCloseFrameIsNotSent()`

[trustin]

Source

WebSocketServiceHttp2RequestCompleteTest > closeAfterTimeoutWhenCloseFrameIsNotSent() FAILED
    org.awaitility.core.ConditionTimeoutException: Condition was evaluated in 100 milliseconds which is earlier than expected minimum timeout 1 seconds
        at app//org.awaitility.core.ConditionAwaiter.await(ConditionAwaiter.java:172)
        at app//org.awaitility.core.CallableCondition.await(CallableCondition.java:78)
        at app//org.awaitility.core.CallableCondition.await(CallableCondition.java:26)
        at app//org.awaitility.core.ConditionFactory.until(ConditionFactory.java:1160)
        at app//org.awaitility.core.ConditionFactory.until(ConditionFactory.java:1129)
        at app//com.linecorp.armeria.server.websocket.WebSocketServiceHttp2RequestCompleteTest.closeAfterTimeoutWhenCloseFrameIsNotSent(WebSocketServiceHttp2RequestCompleteTest.java:124)

[kwondh5217]

Hi, @trustin !

I investigated this issue and think I may have found a possible reason.

It seems there could be a timing issue between when the server accepts the WebSocket connection and when the client’s .thenAccept(...) callback runs. Because of that, updateRequest(...) might not be called in time, which could lead to the connection being closed earlier than expected and the test failing.

Would it make sense to try updating the request directly instead of relying on the .thenAccept(...) callback?

final AbstractWebSocketHandler handler = new AbstractWebSocketHandler() {
    @Override
    void onOpen(WebSocketWriter writer) {
        isConnected.set(true);
        writer.close();
    }
};

...
// busy waiting
new Thread(() -> {
    while (true) {
        if (isConnected.get()) {
            captor.get().updateRequest(HttpRequest.of(HttpMethod.GET, "/"));
            break;
        }
    }
}).start();

I'd love to hear your thoughts !

[minwoox]

@kwondh5217 Thanks for the analysis. I think we can simply fix that by setting requestAutoAbortDelayMillis to the webClient instead of calling updateRequest. Let me fix that soon. 😉