IV22010: LINUX OS AGENT TERMINATES DUE TO SIGHUP SIGNAL

IBM Tivoli Monitoring 6.2.3 Fix Pack 5 (6.2.3-TIV-ITM-FP0005)
IBM Tivoli Monitoring 6.2.3 Fix Pack 2 (6.2.3-TIV-ITM-FP0002)
IBM Tivoli Monitoring 6.2.3 Fix Pack 4 (6.2.3-TIV-ITM-FP0004)
IBM Tivoli Monitoring: Linux (R) OS Agent 6.2.3.1-TIV-ITM_LINUX-IF0001

APAR status

• Closed as program error.

Error description

• Problem Description: the Monitoring Agent for Linux OS
  terminates due to SIGHUP signal if agent is started using the
  following command with -t option:
        ssh -t -p 22 root@XXXX "/etc/init.d/ITMAgents1 stop &&
        /etc/init.d/ITMAgents1 start && /opt/IBM/ITM/bin/cinfo -r"
  
  NOTE:
  This issue is NOT specific to starting the agent through a
  remote session using "ssh".
  This issue is NOT specific to using the -t option on the
  start command.
  
  The LZ agent can also terminate due to a SIGHUP signal
  being received on reboot, which will result in a
  termination of the Linux OS agent with no indication
  in RAS1 logs or other diagnostic logs that a signal has
  been received.
  The LZ agent CAN be started successfully and remain
  running when running either the autostart script after
  the OS initializes (ITMAgents#), or running
  "itmcmd agent start lz".
  
  There will be NO tracepoint from "signalmanager.cpp"
  indicating that the agent received a termination signal.
  
  The LZ agent will only run for a few seconds before
  the LZ agent shuts down.
  
  The .LG0 and RAS1 logs will show what appears to be a
  controlled shutdown / termination:
  In the .LG0:
  KRARMN000 IRA rmt server shutting down.
  
  In RAS1 logging:
  -1:kraafmn.cpp,384,"kramain") Exit: 0x0
  -1:klzmain.cpp,458,"CtMain") kramain returned
  -1:klzmain.cpp,459,"CtMain") Exit: 0x0
  -1:klzmain.cpp,102,"stop_exit") Active RAS1 Classes: EVERYT
  EVERYE EVERYU
  -1:klzmain.cpp,102,"stop_exit") Entry
  -1:kraafmgr.cpp,2291,"IRA_Shutdown") Active RAS1 Classes: EVERYT
  EVERYE EVERYU
  -1:kraafmgr.cpp,2291,"IRA_Shutdown") Entry
  -1:kraafmgr.cpp,2293,"IRA_Shutdown") Shutting down agent remote
  manager.
  
  Reproducing to set KBB_SIG1=-dumpoff will NOT result
  in a core being generated.
  
  Code analysis had to be performed to determine that the
  reason the agent was shutting down was due to the
  rpc__listen() thread exiting.
  
  The issue is NOT readily reproducible in lab testing.
  A diagnostic module had to be built to gather further
  information as well as concurrent "strace" outputs
  to understand what was happening when the
  rpc__listen() threads were exiting on the start
  attempt during reboot.
  
  The strace was necessary to confirm that a SIGHUP was
  being received.
  This APAR is also available in an interim fix based on
  6.22 FP5 level:
  6.2.2.5-TIV-ITM_LINUX-IF0001
  

Local fix

• Do not use -t option (only applies when using "ssh")
  

Problem summary

• Linux OS Agent terminates due to sighup signal.
  
  
  The Monitoring Agent for Linux OS terminates due to unhandled
  SIGHUP signal if agent is started remotely using ssh -t as
  follows:
  
  ssh -t -p 22 root@XXXX "/etc/init.d/ITMAgents1 stop &&
  /etc/init.d/ITMAgents1 start && /opt/IBM/ITM/bin/cinfo -r"
  

Problem conclusion

• The code was changed to handle the signal.
  
  The fix for this APAR is going to be included in the following
  maintenance vehicle:
      | fix pack | 6.2.3-TIV-ITM-FP0002
  

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

  ITM AGENT LINUX

• Fixed component ID

  5724C04LN

Applicable component levels

• R622 PSY

     UP

• R623 PSY

     UP