Sorry, I've been tricked by the client logs which didn't log the reconnection attempts and ping probes (for some reason I remember they should have been logged).
So everything works as expected, it was just a network timing problem. Cheers, Sergio B. On Mon, Jan 18, 2010 at 10:33 AM, Sergio Bossa <sergio.bo...@gmail.com> wrote: > Hi guys, > > I'm having problems with the l1.healthcheck.l2.socketConnect enabled > (as by default). > If I disconnect L1 from its L2 by taking down the network route, I get > the following error message on L2: > > 2010-01-18 09:36:51,919 [HealthChecker] INFO > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl. DSO Server > - Socket Connect to 10.5.10.129:61615(callbackport:61610) failed: > java.net.NoRouteToHostException: No route to host > 2010-01-18 09:36:51,919 [HealthChecker] INFO > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl. DSO Server > - 10.5.10.129:61615 is DEAD > 2010-01-18 09:36:51,919 [HealthChecker] ERROR > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl: DSO Server > - Declared connection dead > ConnectionID(2.735c805963d34b87a9bee2eeaa8bc983) idle time 8476ms > > And the following on L1: > > 2010-01-18 09:36:51,444 [HealthChecker] INFO > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl. DSO Client > - Socket Connect to 10.5.10.129:9510(callbackport:9510) failed: > java.net.NoRouteToHostException: No route to host > 2010-01-18 09:36:51,445 [HealthChecker] INFO > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl. DSO Client > - 10.5.10.129:9510 is DEAD > 2010-01-18 09:36:51,445 [HealthChecker] ERROR > com.tc.net.protocol.transport.ConnectionHealthCheckerImpl: DSO Client > - Declared connection dead > ConnectionID(2.735c805963d34b87a9bee2eeaa8bc983) idle time 8002ms > > However, despite having enabled the "l2.l1reconnect.enabled" property, > reconnection doesn't actually happen when the network route is > reestablished, apparently because the L1 suddenly changes its session > state: > > 2010-01-18 09:36:51,446 [L1_L2:TCComm Main Selector Thread (listen > 0.0.0.0:61610)] INFO com.tc.net.protocol.tcm.ClientMessageChannel - > ClientMessageChannel moves to SessionID=[1] > > As you can see by looking at the timestamps, that happens *prior* to > attempting any reconnection. > > If I disable the "l1.healthcheck.l2.socketConnect" property, > everything works as expected, and the L1 attempts to reconnect prior > to moving session state. > > Any clues? > You can find client and server logs in attachment. > > Cheers, > > Sergio B. > > -- > Sergio Bossa > Software Passionate and Open Source Enthusiast. > URL: http://www.linkedin.com/in/sergiob > -- Sergio Bossa Software Passionate and Open Source Enthusiast. URL: http://www.linkedin.com/in/sergiob _______________________________________________ tc-dev mailing list tc-dev@lists.terracotta.org http://lists.terracotta.org/mailman/listinfo/tc-dev
