IV60096: SOCKET LEAK WHEN DIRECTORY POINTED TO BY KCA_IP_DIR VARIABLE DOES NOT EXIST

APAR status

• Closed as program error.

Error description

• Approver:         MK
  Environment:      ITM 623
  
  Problem Description:  The OS Agent communicates with the
  Watchdog process (kcawd) via a directory defined by KCA_IP_DIR.
  
  The default for this directory is
  $CANDLEHOME$/$BINARCH$/$PRODUCTCODE$/bin/pasipc.   If this value
  is overwritten and/or does not exist it will result in a socket
  leak.
  
  The RAS1 OS Agent log file with tracing set
  to (UNIT:KCA ALL) will show the messages similar to what is
  shown below every two mintues:
  
  --------------
  (535F5D3F.0045-4:kcasockclient.cpp,163,"sendCommand") Attempting
  to connect to socket /opt/IBM/ITM/kca/.pas_wd_sock.
  (535F5D3F.0046-4:kcasockclient.cpp,167,"sendCommand") Error
  executing connect - errno = 2, No such file or directory
  --------------
  
  Running "netstat -an" will show several sockets opened. The
  number of sockets open will come back to normal if the Unix
  agent is stopped or if watchdog is disabled.
  
  Affected levels: 6.23.xx.xx and 6.30 base for the UNIX and Linux
  OS Agents.
  
  Keywords:
  watchdog, agent management services, AMS, socket, kcawd
  

Local fix

• Manually create the directory defined by the KCA_IP_DIR
  variable.
  

Problem summary

• When the KCA_IP_DIR environment variable value is changed from
  the default value of
  CANDLEHOME$/$BINARCH$/$PRODUCTCODE$/bin/pasipc and the new
  directory specified does not exist, the code does not close the
  socket when trying to communicate with the Watchdog (kcawd)
  process.   This results in a socket leak.
  
  
  If the value for the KCA_IP_DIR is changed from the default
  value, the directory must exist and be writeable by the OS
  Agent. If the directory does not exist, the following messages
  will be seen in the Monitoring Agent for UNIX OS and Watchdog
  RAS1 log files every status check.
  
  
  (53B408FB.0000-A:kcasockclient.cpp,167,"sendCommand") Error
  executing
  connect - errno = 2, No such file or directory
  (53B408FB.0001-A:kcawd.cpp,167,"getITMAvailabilityOnPipe") ***
  ERROR: rc = 3,
  executing GET_ID query on WatchDog Agent;
  +53B408FB.0001 command: <622 QUERY GET_PID>
  +53B408FB.0001 result: <>
  
  Regular status checks are first attempted using this connection
  and if that is not successful will check status two other ways,
  so the messages will appear in the log but will not hinder the
  operation of the Watchdog function.
  
  
  The following functions also depend on the existence of the
  directory specified by the KCA_IP_DIR:
  - The reporting of the OS Agent restarts to the Agent Management
    Services workspace.
  - The task to recycle the Monitoring Agent for UNIX OS
  

Problem conclusion

• The code was updated to close the socket on the error condition.
  
  
  The fix for this APAR is contained in the following maintenance
  packages:
  
    | fix pack | 6.3.0-TIV-ITM-FP0001
  

Temporary fix

• Manually create the directory defined by the KCA_IP_DIR
  variable.
  

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 UNIX

• Fixed component ID

  5724C040U

Applicable component levels

• R630 PSY

     UP