[GEODE-8005] NPE on locator when connecting to ClusterManagementService using a cache.

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: gfsh, management
• Assignee: Unassigned
• Reporter: Patrick Johnsn
• Created: 20/Apr/20 18:18
• Updated: 20/Apr/20 18:18

Description

Starting a cluster in gfsh like this:

gfsh> start locator
gfsh> start server

And creating a GeodeConnectionConfig like this:

 ClientCache clientCache = new ClientCacheFactory()
 .addPoolLocator("localhost", 10334)
 .create();

ConnectionConfig connectionConfig = new GeodeConnectionConfig(clientCache);

will result in the following exception on the client:

Exception in thread "main" java.lang.IllegalStateException: Unable to discover a locator that has ClusterManagementService running.Exception in thread "main" java.lang.IllegalStateException: Unable to discover a locator that has ClusterManagementService running. at org.apache.geode.management.internal.api.GeodeConnectionConfig.setClientCache(GeodeConnectionConfig.java:126) at org.apache.geode.management.internal.api.GeodeConnectionConfig.<init>(GeodeConnectionConfig.java:74) at example.v2.Experiment.main(Experiment.java:42) 

 And this exception in the locator logs: 

java.lang.NullPointerExceptionjava.lang.NullPointerException at org.apache.geode.management.internal.configuration.handlers.ClusterManagementServiceInfoRequestHandler.processRequest(ClusterManagementServiceInfoRequestHandler.java:56) at org.apache.geode.distributed.internal.membership.gms.locator.PrimaryHandler.processRequest(PrimaryHandler.java:70) at org.apache.geode.distributed.internal.tcpserver.TcpServer.processOneConnection(TcpServer.java:421) at org.apache.geode.distributed.internal.tcpserver.TcpServer.lambda$processRequest$0(TcpServer.java:326) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) 

A workaround for this is to set "--bind-address=localhost" when starting the locator.

 

Despite this exception, the ClusterManagementService is running on the locator and the endpoints can still be accessed using hostname and port.