You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@geode.apache.org by "Dan Smith (JIRA)" <ji...@apache.org> on 2016/01/26 01:50:39 UTC

[jira] [Created] (GEODE-858) InternalLocator.startTcpServer has a 1 second sleep

Dan Smith created GEODE-858:
-------------------------------

             Summary: InternalLocator.startTcpServer has a 1 second sleep
                 Key: GEODE-858
                 URL: https://issues.apache.org/jira/browse/GEODE-858
             Project: Geode
          Issue Type: Bug
            Reporter: Dan Smith


This method has a 1 second sleep. Looking at the code, there doesn't seem to be any reason to sleep. server.start has already created the tcp socket for the server, so clients should be able to connect after that call has done. They will wait until the acceptor thread invokes accept, but that seems reasonable. 

The issue with this sleep is that all of our WAN tests (several hundred) start multiple locators in each test, so this adds a lot of time to our WAN tests.

{code}
private void startTcpServer() throws IOException {
    logger.info(LocalizedMessage.create(LocalizedStrings.InternalLocator_STARTING_0, this));
    server.start();
    
    try { 
      Thread.sleep(1000); 
    } 
    catch (InterruptedException ie) {
      // always safe to exit this thread...
      Thread.currentThread().interrupt();
      logger.warn(LocalizedMessage.create(LocalizedStrings.ONE_ARG, "Interrupted"), ie);
    }
  }
{code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)