IV70977: INTERMITTENTLY RDMA ENDPOINT CONNECT WAITS FOREVER

APAR status

• Closed as program error.

Error description

• Error Message: RDMA Endpoint connect waits forever without event
  in multithreaded scenario. It happens intermittently.
  .
  Stack Trace: Using "kill -3", we can get below stack which shows
  the issueXMTHREADINFO      "pool-2-thread-7"
  J9VMThread:0x00007FDCECC4FE00, j9thread_t:0x00007FDCECC4BB70,
  java/lang/Thread:0x00007FDCD406B478, state:R,
  prio=53XMJAVALTHREAD            (java/lang/Thread getId:0x35,
  isDaemon:false)3XMTHREADINFO1            (native thread
  ID:0x1A77, native priority:0x5, native policy:UNKNOWN,
  vmstate:R, vm thread flags:0x00000001)3XMTHREADINFO2
  (native stack address range from:0x00007FDCB0DCE000,
  to:0x00007FDCB0E0F000, size:0x41000)3XMCPUTIME               CPU
  usage total: 135.820678986 secs, current
  category="Application"3XMHEAPALLOC             Heap bytes
  allocated since last GC cycle=0 (0x0)3XMTHREADINFO3
  Java callstack:4XESTACKTRACE                at
  com/ibm/net/rdma/jverbs/verbs/NativeRdmaJVerbs.getCqEvent0(Nativ
  e Method)4XESTACKTRACE                at
  com/ibm/net/rdma/jverbs/verbs/NativeRdmaJVerbs.getCQEvent(Native
  RdmaJVerbs.java:174(Compiled Code))4XESTACKTRACE
  at
  com/ibm/net/rdma/jverbs/verbs/CompletionChannel.getCQEvent(Compl
  etionChannel.java:75(Compiled Code))4XESTACKTRACE
  at
  com/ibm/jverbs/test/util/RdmaDataPath.checkCq(RdmaDataPath.java:
  165(Compiled Code))4XESTACKTRACE                at
  com/ibm/jverbs/test/util/RdmaDataPath.completeSGRecv(RdmaDataPat
  h.java:104(Compiled Code))4XESTACKTRACE                at
  com/ibm/jverbs/test/verbs/client/JverbsRdmaClient.verbs_repeated
  SendRecv(JverbsRdmaClient.java:666)4XESTACKTRACE
  at sun/reflect/GeneratedMethodAccessor10.invoke(Bytecode
  PC:36)4XESTACKTRACE                at
  sun/reflect/DelegatingMethodAccessorImpl.invoke(DelegatingMethod
  AccessorImpl.java:55(Compiled Code))4XESTACKTRACE
  at java/lang/reflect/Method.invoke(Method.java:495(Compiled
  Code))4XESTACKTRACE                at
  org/testng/internal/MethodInvocationHelper.invokeMethod(MethodIn
  vocationHelper.java:80)4XESTACKTRACE                at
  org/testng/internal/Invoker.invokeMethod(Invoker.java:702)4XESTA
  CKTRACE                at
  org/testng/internal/Invoker.invokeTestMethod(Invoker.java:894)4X
  ESTACKTRACE                at
  org/testng/internal/Invoker.invokeTestMethods(Invoker.java:1219)
  4XESTACKTRACE                at
  org/testng/internal/TestMethodWorker.invokeTestMethods(TestMetho
  dWorker.java:127)4XESTACKTRACE                at
  org/testng/internal/TestMethodWorker.run(TestMethodWorker.java:1
  11)4XESTACKTRACE                at
  java/util/concurrent/ThreadPoolExecutor.runWorker(ThreadPoolExec
  utor.java:1153(Compiled Code))4XESTACKTRACE                at
  java/util/concurrent/ThreadPoolExecutor$Worker.run(ThreadPoolExe
  cutor.java:628)4XESTACKTRACE                at
  java/lang/Thread.run(Thread.java:785)
  .
  

Local fix

Problem summary

• RDMA Endpoint connect waits forever due to RDMA CONNECTION event
  misses.
  

Problem conclusion

• Corrected event misses which is due to unexpected error.
  Tightened the exception handling.
  .
  This APAR will be fixed in the following Java Releases:
     8    SR1       (8.0.1.0)
     7 R1 SR3       (7.1.3.0)
     7    SR9       (7.0.9.0)
  .
  Contact your IBM Product's Service Team for these Service
  Refreshes and Fix Packs.
  For those running stand-alone, information about the available
  Service Refreshes and Fix Packs can be found at:
             https://www.ibm.com/developerworks/java/jdk/
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  JAVA CLASS LIBS

• Fixed component ID

  620700130

Applicable component levels

• R800 PSY

     UP

• R700 PSY

     UP