[IGNITE-11671] Thin client: Client may hang when connected to a starting server

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.8
• Component/s: thin client
• Assignee: Aleksey Plekhanov
• Reporter: Aleksey Plekhanov
• Created: 02/Apr/19 13:10
• Updated: 04/Sep/19 18:19
• Resolved: 23/May/19 14:06

Description

If the server start process has not completed yet, but NIO listeners already started, the client may never get a response for the handshake request.

Exception on the server-side:

 

[client-connector-#6416%f3b837aa-d726-46b0-a58b-8cc6267c9f96%][ClientListenerProcessor] Runtime error caught during grid runnable execution: GridWorker [name=message-received-notify, igniteInstanceName=f3b837aa-d726-46b0-a58b-8cc6267c9f96, finished=false, heartbeatTs=1554209548706, hashCode=519781823, interrupted=false, runner=client-connector-#6416%f3b837aa-d726-46b0-a58b-8cc6267c9f96%]
java.lang.NullPointerException
at org.apache.ignite.internal.processors.odbc.ClientListenerNioListener.nextConnectionId(ClientListenerNioListener.java:334)
at org.apache.ignite.internal.processors.odbc.ClientListenerNioListener.prepareContext(ClientListenerNioListener.java:313)
at org.apache.ignite.internal.processors.odbc.ClientListenerNioListener.onHandshake(ClientListenerNioListener.java:251)
at org.apache.ignite.internal.processors.odbc.ClientListenerNioListener.onMessage(ClientListenerNioListener.java:132)
at org.apache.ignite.internal.processors.odbc.ClientListenerNioListener.onMessage(ClientListenerNioListener.java:48)
at org.apache.ignite.internal.util.nio.GridNioFilterChain$TailFilter.onMessageReceived(GridNioFilterChain.java:279)
at org.apache.ignite.internal.util.nio.GridNioFilterAdapter.proceedMessageReceived(GridNioFilterAdapter.java:109)
at org.apache.ignite.internal.util.nio.GridNioAsyncNotifyFilter$3.body(GridNioAsyncNotifyFilter.java:97)
at org.apache.ignite.internal.util.worker.GridWorker.run(GridWorker.java:120)
at org.apache.ignite.internal.util.worker.GridWorkerPool$1.run(GridWorkerPool.java:70)

 

This happens because NIO listeners start before GridDiscoveryManager.

Issue Links

• is duplicated by: IGNITE-12137 ODBC: Application hang when connecting to a starting single node
• links to: GitHub Pull Request #6499
• links to: GitHub Pull Request #6550
• links to: TC
• links to: Upsource