0 Replies Latest reply on Nov 20, 2018 4:33 AM by Thomas Houtekier

    initial state transfer timed out while it is disabled

    Thomas Houtekier Newbie

      I run into this problem at startup with wildfly-9.0.2.Final which contains Infinispan-7.2.3.Final.

      At startup, the cache throws a org.infinispan.commons.CacheException: Initial state transfer timed out for cache myCache

       

      Caused by: org.infinispan.commons.CacheException: Initial state transfer timed out for cache myCache on 192.168.7.2
      at org.infinispan.statetransfer.StateTransferManagerImpl.waitForInitialStateTransferToComplete(StateTransferManagerImpl.java:222) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) [rt.jar:1.8.0_181]
      at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source) [rt.jar:1.8.0_181]
      at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source) [rt.jar:1.8.0_181]
      at java.lang.reflect.Method.invoke(Unknown Source) [rt.jar:1.8.0_181]
      at org.infinispan.commons.util.ReflectionUtil.invokeAccessibly(ReflectionUtil.java:168)
      ... 91 more

      It is thrown when requesting the cache from the cachecontainer:

       

       

      at org.infinispan.commons.util.ReflectionUtil.invokeAccessibly(ReflectionUtil.java:171)
      at org.infinispan.factories.AbstractComponentRegistry$PrioritizedMethod.invoke(AbstractComponentRegistry.java:869) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.factories.AbstractComponentRegistry.invokeStartMethods(AbstractComponentRegistry.java:638) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.factories.AbstractComponentRegistry.internalStart(AbstractComponentRegistry.java:627) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.factories.AbstractComponentRegistry.start(AbstractComponentRegistry.java:530) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.factories.ComponentRegistry.start(ComponentRegistry.java:218) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.cache.impl.CacheImpl.start(CacheImpl.java:850) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.manager.DefaultCacheManager.wireAndStartCache(DefaultCacheManager.java:599) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.manager.DefaultCacheManager.createCache(DefaultCacheManager.java:554) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.manager.DefaultCacheManager.getCache(DefaultCacheManager.java:424) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.infinispan.manager.DefaultCacheManager.getCache(DefaultCacheManager.java:438) [infinispan-core-7.2.3.Final.jar:7.2.3.Final]
      at org.jboss.as.clustering.infinispan.DefaultCacheContainer.getCache(DefaultCacheContainer.java:118)
      at org.jboss.as.clustering.infinispan.DefaultCacheContainer.getCache(DefaultCacheContainer.java:109)

       

      The remarkable thing is that initial-statetransfer is disabled by configuration for this cache. The reason it is disabled is to prevent this exact exception which I've seen before.

      I don't really care about the initial-statetransfer, that is why it is disabled. Since it is disabled, I've not seen this exception again, until now.

      This is the config:

       

      <cache-container name="interserver" default-cache="default" jndi-name="java:jboss/infinispan/interserver" >
      <transport stack="udp-interserver" lock-timeout="60000" />
      <replicated-cache name="default" mode="SYNC">
      <locking isolation="READ_COMMITTED"/>
      <transaction mode="BATCH"/>
      <state-transfer enabled="false"/>
      </replicated-cache>
      <replicated-cache name="myCache"  mode="ASYNC">
      <locking isolation="READ_COMMITTED"/>
      <transaction mode="BATCH"/>
      <state-transfer enabled="false"/>                    
      </replicated-cache>
      </cache-container>                    
      

       

       

      What changed is that now I start multiple nodes (14) at the same time. This is something that didn't happen before.

      Is there anything special with starting multiple nodes at the same time which can explain this?

      I see this exception quite regularly on a few of the nodes that are starting. When restarting the server, it typically resolves.

      There are also other exceptions that sometimes happen during startup: NullpointerException when starting Infinispan

      How come this exception is thrown, while the initial-state-transfer is disabled?