PH45851: INTEGRATION SERVERS DID NOT SHUT DOWN DURING IPL

APAR status

• Closed as documentation error.

Error description

• A deadlock situation on the ConfigurationMutex lock results in
  an endless loop waiting for it to be released so that the
  ImbStatsCollector threadpool can shutdown. A dump of the hung
  Integration Server address space will contain the following two
  thread stacks:
  
  Thread 0:
  
  poll
  imbSleep(int)
  ImbStatsCollector::¯ImbStatsCollector()
  ImbMain::start(ImbStartParameters&)
  main
  
  
  Thread 2:
  
  CEEOPCW
  pthread_cond_wait
  pthread_cond_wait_03
  ImbMutex::lock(int)
  ImbExecutionGroupDirector::returnTheSMFTToPool(ImbDataFlowSM
  ImbStatsCollectorState_Collecting::runState(ImbStatsCollecto
  ImbStatsCollector::runCollectorCycle(ImbOsThread*)
  ImbStatsCollector::Parameters::run(ImbOsThread*)
  ImbThreadPoolThreadFunction::run(ImbOsThread*)
  ImbOsThread::threadRun()
  ImbOsThread::innerThreadBootStrapWrapper(void*)
  ImbOsThread::threadBootStrap(void*)
  threadBootStrapWrapper
  CELQPCMM
  

Local fix

• N/A
  

Problem summary

• ****************************************************************
  USERS AFFECTED:
  All IBM Integration Bus for z/OS 10.1 users on z/OS trying to
  initiate an IPL without stopping the broker first.
  
  
  Platforms affected:
  ZOS
  
  ****************************************************************
  PROBLEM DESCRIPTION:
  A deadlock situation occurs when an IPL is initiated on a zOS
  system while the broker is running.
  

Problem conclusion

• Product Documentation is being updated.
  
  A new procedure,
  explaining how to deal with the situation, will be added to this
  page:
  https://www.ibm.com/docs/en/integration-bus/10.1.0?topic=p
  roblems-resolving-when-stopping-resources.
  
  Title: Execution
  groups stay alive while broker task is killed during an
  IPL.
  
  Scenario: During an IPL, broker task is killed but some
  execution groups are staying alive.
  
  Explanation: Execution
  groups were not stopped as the broker was not stopped normally,
  using the mqsistop command.
  
  Solution: Use the "mqsistart"
  command to start the broker, stop the broker using "mqsistop".
  Then the IPL/Maintenance event could proceed as expected.
  Alternatively, mqsitop command can be used to stop the broker
  before IPL.
  

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

Applicable component levels